id	sid	tid	token	lemma	pos
work_fklnz2kdpjhwfjysurpx7koiti	1	1	Transactions	transaction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	1	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	1	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	1	4	Association	Association	NNP
work_fklnz2kdpjhwfjysurpx7koiti	1	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	1	6	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	1	7	Linguistics	Linguistics	NNP
work_fklnz2kdpjhwfjysurpx7koiti	1	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	1	9	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	1	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	1	11	2013	2013	CD
work_fklnz2kdpjhwfjysurpx7koiti	1	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	1	13	279–290	279–290	CD
work_fklnz2kdpjhwfjysurpx7koiti	1	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	2	1	Action	Action	NNP
work_fklnz2kdpjhwfjysurpx7koiti	2	2	Editor	Editor	NNP
work_fklnz2kdpjhwfjysurpx7koiti	2	3	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	2	4	Lillian	Lillian	NNP
work_fklnz2kdpjhwfjysurpx7koiti	2	5	Lee	Lee	NNP
work_fklnz2kdpjhwfjysurpx7koiti	2	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	3	1	Submitted	submit	VBN
work_fklnz2kdpjhwfjysurpx7koiti	3	2	11/2012	11/2012	CD
work_fklnz2kdpjhwfjysurpx7koiti	3	3	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	3	4	Revised	revise	VBN
work_fklnz2kdpjhwfjysurpx7koiti	3	5	1/2013	1/2013	CD
work_fklnz2kdpjhwfjysurpx7koiti	3	6	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	3	7	Published	publish	VBN
work_fklnz2kdpjhwfjysurpx7koiti	3	8	7/2013	7/2013	NNP
work_fklnz2kdpjhwfjysurpx7koiti	3	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	4	1	c	c	NNP
work_fklnz2kdpjhwfjysurpx7koiti	4	2	©	©	NNP
work_fklnz2kdpjhwfjysurpx7koiti	4	3	2013	2013	CD
work_fklnz2kdpjhwfjysurpx7koiti	4	4	Association	Association	NNP
work_fklnz2kdpjhwfjysurpx7koiti	4	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	4	6	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	4	7	Linguistics	Linguistics	NNP
work_fklnz2kdpjhwfjysurpx7koiti	4	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	5	1	Good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	3	Great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	5	Excellent	excellent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	6	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	5	7	Global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	8	Inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	5	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	5	10	Semantic	Semantic	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	11	Intensities	Intensities	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	5	12	Gerard	Gerard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	13	de	de	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	14	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	15	ICSI	ICSI	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	17	Berkeley	Berkeley	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	18	demelo@icsi.berkeley.edu	demelo@icsi.berkeley.edu	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	19	Mohit	Mohit	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	20	Bansal	Bansal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	21	CS	CS	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	22	Division	Division	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	24	UC	UC	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	25	Berkeley	Berkeley	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	26	mbansal@cs.berkeley.edu	mbansal@cs.berkeley.edu	NN
work_fklnz2kdpjhwfjysurpx7koiti	5	27	Abstract	Abstract	NNP
work_fklnz2kdpjhwfjysurpx7koiti	5	28	Adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	5	29	like	like	IN
work_fklnz2kdpjhwfjysurpx7koiti	5	30	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	32	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	34	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	5	35	excellent	excellent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	36	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	5	37	similar	similar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	5	38	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	5	39	meaning	meaning	NN
work_fklnz2kdpjhwfjysurpx7koiti	5	40	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	5	41	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	5	42	differ	differ	VB
work_fklnz2kdpjhwfjysurpx7koiti	5	43	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	5	44	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	5	45	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	6	1	In-	in-	IN
work_fklnz2kdpjhwfjysurpx7koiti	6	2	tensity	tensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	6	3	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	6	4	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	6	5	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	6	6	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	6	7	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	6	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	6	9	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	6	10	learners	learner	NNS
work_fklnz2kdpjhwfjysurpx7koiti	6	11	as	as	RB
work_fklnz2kdpjhwfjysurpx7koiti	6	12	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	6	13	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	6	14	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	6	15	several	several	JJ
work_fklnz2kdpjhwfjysurpx7koiti	6	16	NLP	NLP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	6	17	tasks	task	NNS
work_fklnz2kdpjhwfjysurpx7koiti	6	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	6	19	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	6	20	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	6	21	missing	miss	VBG
work_fklnz2kdpjhwfjysurpx7koiti	6	22	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	6	23	most	most	JJS
work_fklnz2kdpjhwfjysurpx7koiti	6	24	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	6	25	resources	resource	NNS
work_fklnz2kdpjhwfjysurpx7koiti	6	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	6	27	dictionaries	dictionary	NNS
work_fklnz2kdpjhwfjysurpx7koiti	6	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	6	29	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	6	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	6	31	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	6	32	thesauri	thesauri	NNP
work_fklnz2kdpjhwfjysurpx7koiti	6	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	6	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	7	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	7	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	7	3	paper	paper	NN
work_fklnz2kdpjhwfjysurpx7koiti	7	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	7	5	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	7	6	present	present	VBP
work_fklnz2kdpjhwfjysurpx7koiti	7	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	7	8	primarily	primarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	7	9	unsupervised	unsupervise	VBN
work_fklnz2kdpjhwfjysurpx7koiti	7	10	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	7	11	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	7	12	uses	use	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	7	13	semantics	semantic	NNS
work_fklnz2kdpjhwfjysurpx7koiti	7	14	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	7	15	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	7	16	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	7	17	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	7	18	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	7	19	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	7	20	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	7	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	7	22	phrases	phrase	NNS
work_fklnz2kdpjhwfjysurpx7koiti	7	23	like	like	UH
work_fklnz2kdpjhwfjysurpx7koiti	7	24	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	7	25	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	7	26	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	7	27	excel-	excel-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	7	28	lent	lent	NNP
work_fklnz2kdpjhwfjysurpx7koiti	7	29	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	7	30	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	7	31	rank	rank	VB
work_fklnz2kdpjhwfjysurpx7koiti	7	32	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	7	33	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	7	34	assigning	assign	VBG
work_fklnz2kdpjhwfjysurpx7koiti	7	35	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	7	36	posi-	posi-	VBD
work_fklnz2kdpjhwfjysurpx7koiti	7	37	tions	tion	NNS
work_fklnz2kdpjhwfjysurpx7koiti	7	38	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	7	39	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	7	40	continuous	continuous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	7	41	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	7	42	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	8	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	8	2	rely	rely	VBP
work_fklnz2kdpjhwfjysurpx7koiti	8	3	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	8	4	Mixed	Mixed	NNP
work_fklnz2kdpjhwfjysurpx7koiti	8	5	Integer	Integer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	8	6	Linear	Linear	NNP
work_fklnz2kdpjhwfjysurpx7koiti	8	7	Programming	Programming	NNP
work_fklnz2kdpjhwfjysurpx7koiti	8	8	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	8	9	jointly	jointly	RB
work_fklnz2kdpjhwfjysurpx7koiti	8	10	deter-	deter-	VB
work_fklnz2kdpjhwfjysurpx7koiti	8	11	mine	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	8	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	8	13	ranks	rank	NNS
work_fklnz2kdpjhwfjysurpx7koiti	8	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	8	15	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	8	16	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	8	17	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	8	18	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	8	19	benefit	benefit	VBP
work_fklnz2kdpjhwfjysurpx7koiti	8	20	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	8	21	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	8	22	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	8	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	9	1	When	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	9	2	rank-	rank-	RB
work_fklnz2kdpjhwfjysurpx7koiti	9	3	ing	e	VBG
work_fklnz2kdpjhwfjysurpx7koiti	9	4	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	9	5	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	9	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	9	7	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	9	8	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	9	9	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	10	achieves	achieve	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	9	11	substantial	substantial	JJ
work_fklnz2kdpjhwfjysurpx7koiti	9	12	improvements	improvement	NNS
work_fklnz2kdpjhwfjysurpx7koiti	9	13	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	9	14	pre-	pre-	DT
work_fklnz2kdpjhwfjysurpx7koiti	9	15	vious	vious	JJ
work_fklnz2kdpjhwfjysurpx7koiti	9	16	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	17	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	9	18	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	9	19	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	20	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	9	21	rank	rank	NNP
work_fklnz2kdpjhwfjysurpx7koiti	9	22	corre-	corre-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	9	23	lation	lation	NNP
work_fklnz2kdpjhwfjysurpx7koiti	9	24	metrics	metric	NNS
work_fklnz2kdpjhwfjysurpx7koiti	9	25	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	9	26	specifically	specifically	RB
work_fklnz2kdpjhwfjysurpx7koiti	9	27	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	9	28	70	70	CD
work_fklnz2kdpjhwfjysurpx7koiti	9	29	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	30	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	31	ac-	ac-	DT
work_fklnz2kdpjhwfjysurpx7koiti	9	32	curacy	curacy	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	33	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	9	34	compared	compare	VBN
work_fklnz2kdpjhwfjysurpx7koiti	9	35	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	9	36	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	9	37	56	56	CD
work_fklnz2kdpjhwfjysurpx7koiti	9	38	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	39	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	9	40	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	9	41	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	9	42	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	9	43	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	10	1	Moreover	moreover	RB
work_fklnz2kdpjhwfjysurpx7koiti	10	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	10	3	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	10	4	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	10	5	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	10	6	incorpo-	incorpo-	VB
work_fklnz2kdpjhwfjysurpx7koiti	10	7	rate	rate	VB
work_fklnz2kdpjhwfjysurpx7koiti	10	8	external	external	JJ
work_fklnz2kdpjhwfjysurpx7koiti	10	9	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	10	10	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	10	11	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	10	12	increas-	increas-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	10	13	ing	ing	NNP
work_fklnz2kdpjhwfjysurpx7koiti	10	14	its	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	10	15	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	10	16	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	10	17	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	10	18	78	78	CD
work_fklnz2kdpjhwfjysurpx7koiti	10	19	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	10	20	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	10	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	10	22	extends	extend	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	10	23	easily	easily	RB
work_fklnz2kdpjhwfjysurpx7koiti	10	24	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	10	25	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	10	26	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	10	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	11	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	11	2	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	11	3	make	make	VBP
work_fklnz2kdpjhwfjysurpx7koiti	11	4	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	11	5	code	code	NN
work_fklnz2kdpjhwfjysurpx7koiti	11	6	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	11	7	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	11	8	freely	freely	RB
work_fklnz2kdpjhwfjysurpx7koiti	11	9	available.1	available.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	11	10	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	11	11	Introduction	Introduction	NNP
work_fklnz2kdpjhwfjysurpx7koiti	11	12	Current	current	JJ
work_fklnz2kdpjhwfjysurpx7koiti	11	13	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	11	14	resources	resource	NNS
work_fklnz2kdpjhwfjysurpx7koiti	11	15	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	11	16	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	11	17	dictionaries	dictionary	NNS
work_fklnz2kdpjhwfjysurpx7koiti	11	18	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	11	19	thesauri	thesauri	NNP
work_fklnz2kdpjhwfjysurpx7koiti	11	20	do	do	VBP
work_fklnz2kdpjhwfjysurpx7koiti	11	21	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	11	22	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	11	23	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	11	24	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	11	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	11	26	in-	in-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	11	27	tensity	tensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	11	28	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	11	29	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	11	30	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	11	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	12	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	12	2	example	example	NN
work_fklnz2kdpjhwfjysurpx7koiti	12	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	12	4	both	both	CC
work_fklnz2kdpjhwfjysurpx7koiti	12	5	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	12	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	12	7	Miller	Miller	NNP
work_fklnz2kdpjhwfjysurpx7koiti	12	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	12	9	1995	1995	CD
work_fklnz2kdpjhwfjysurpx7koiti	12	10	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	12	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	12	12	Roget	Roget	NNP
work_fklnz2kdpjhwfjysurpx7koiti	12	13	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	12	14	21st	21st	JJ
work_fklnz2kdpjhwfjysurpx7koiti	12	15	Century	Century	NNP
work_fklnz2kdpjhwfjysurpx7koiti	12	16	Thesaurus	Thesaurus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	12	17	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	12	18	thesaurus.com	thesaurus.com	ADD
work_fklnz2kdpjhwfjysurpx7koiti	12	19	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	12	20	present	present	JJ
work_fklnz2kdpjhwfjysurpx7koiti	12	21	acceptable	acceptable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	12	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	12	23	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	12	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	12	25	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	12	26	su-	su-	RB
work_fklnz2kdpjhwfjysurpx7koiti	12	27	perb	perb	NN
work_fklnz2kdpjhwfjysurpx7koiti	12	28	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	12	29	synonyms	synonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	12	30	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	12	31	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	12	32	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	12	33	good	good	NN
work_fklnz2kdpjhwfjysurpx7koiti	12	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	13	1	However	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	13	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	13	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	13	4	native	native	JJ
work_fklnz2kdpjhwfjysurpx7koiti	13	5	speaker	speaker	NN
work_fklnz2kdpjhwfjysurpx7koiti	13	6	knows	know	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	13	7	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	13	8	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	13	9	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	13	10	represent	represent	VBP
work_fklnz2kdpjhwfjysurpx7koiti	13	11	varying	varying	NN
work_fklnz2kdpjhwfjysurpx7koiti	13	12	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	13	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	13	14	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	13	15	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	13	16	fact	fact	NN
work_fklnz2kdpjhwfjysurpx7koiti	13	17	generally	generally	RB
work_fklnz2kdpjhwfjysurpx7koiti	13	18	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	13	19	ranked	rank	VBN
work_fklnz2kdpjhwfjysurpx7koiti	13	20	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	13	21	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	13	22	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	13	23	acceptable	acceptable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	13	24	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	13	25	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	13	26	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	13	27	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	13	28	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	13	29	superb	superb	JJ
work_fklnz2kdpjhwfjysurpx7koiti	13	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	14	1	Similarly	similarly	RB
work_fklnz2kdpjhwfjysurpx7koiti	14	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	14	3	warm	warm	JJ
work_fklnz2kdpjhwfjysurpx7koiti	14	4	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	14	5	hot	hot	JJ
work_fklnz2kdpjhwfjysurpx7koiti	14	6	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	14	7	scorching	scorching	NN
work_fklnz2kdpjhwfjysurpx7koiti	14	8	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	14	9	identified	identify	VBN
work_fklnz2kdpjhwfjysurpx7koiti	14	10	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	14	11	synonyms	synonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	14	12	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	14	13	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	14	14	resources	resource	NNS
work_fklnz2kdpjhwfjysurpx7koiti	14	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	15	1	Ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	15	2	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	15	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	15	4	1http://demelo.org/gdm/intensity/	1http://demelo.org/gdm/intensity/	NN
work_fklnz2kdpjhwfjysurpx7koiti	15	5	however	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	15	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	15	7	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	15	8	crucial	crucial	JJ
work_fklnz2kdpjhwfjysurpx7koiti	15	9	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	15	10	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	15	11	allows	allow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	15	12	us	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	15	13	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	15	14	differen-	differen-	VB
work_fklnz2kdpjhwfjysurpx7koiti	15	15	tiate	tiate	NN
work_fklnz2kdpjhwfjysurpx7koiti	15	16	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	16	1	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	2	various	various	JJ
work_fklnz2kdpjhwfjysurpx7koiti	16	3	intensities	intensity	NNS
work_fklnz2kdpjhwfjysurpx7koiti	16	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	5	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	16	6	emotion	emotion	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	8	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	9	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	16	10	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	16	11	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	16	12	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	16	13	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	14	humans	human	NNS
work_fklnz2kdpjhwfjysurpx7koiti	16	15	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	16	16	learning	learn	VBG
work_fklnz2kdpjhwfjysurpx7koiti	16	17	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	16	18	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	19	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	20	judging	judge	VBG
work_fklnz2kdpjhwfjysurpx7koiti	16	21	product	product	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	22	reviews	review	NNS
work_fklnz2kdpjhwfjysurpx7koiti	16	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	24	as	as	RB
work_fklnz2kdpjhwfjysurpx7koiti	16	25	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	16	26	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	27	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	28	automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	16	29	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	30	understanding	understanding	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	31	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	32	generation	generation	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	33	tasks	task	NNS
work_fklnz2kdpjhwfjysurpx7koiti	16	34	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	16	35	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	16	36	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	37	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	38	subjectivity	subjectivity	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	39	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	40	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	41	recog-	recog-	VBD
work_fklnz2kdpjhwfjysurpx7koiti	16	42	nizing	nize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	16	43	textual	textual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	16	44	entailment	entailment	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	45	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	46	question	question	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	47	answering	answering	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	48	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	49	sum-	sum-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	16	50	marization	marization	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	51	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	16	52	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	53	coreference	coreference	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	54	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	16	55	discourse	discourse	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	56	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	16	57	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	17	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	17	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	17	3	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	17	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	17	5	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	17	6	attempt	attempt	VBP
work_fklnz2kdpjhwfjysurpx7koiti	17	7	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	17	8	automatically	automatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	17	9	rank	rank	VB
work_fklnz2kdpjhwfjysurpx7koiti	17	10	sets	set	NNS
work_fklnz2kdpjhwfjysurpx7koiti	17	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	17	12	related	related	JJ
work_fklnz2kdpjhwfjysurpx7koiti	17	13	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	17	14	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	17	15	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	17	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	17	17	focusing	focus	VBG
work_fklnz2kdpjhwfjysurpx7koiti	17	18	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	17	19	par-	par-	NN
work_fklnz2kdpjhwfjysurpx7koiti	17	20	ticular	ticular	NN
work_fklnz2kdpjhwfjysurpx7koiti	17	21	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	17	22	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	17	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	18	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	18	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	18	3	made	make	VBN
work_fklnz2kdpjhwfjysurpx7koiti	18	4	possible	possible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	18	5	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	18	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	18	7	vast	vast	JJ
work_fklnz2kdpjhwfjysurpx7koiti	18	8	amounts	amount	NNS
work_fklnz2kdpjhwfjysurpx7koiti	18	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	18	10	world	world	NN
work_fklnz2kdpjhwfjysurpx7koiti	18	11	knowledge	knowledge	NN
work_fklnz2kdpjhwfjysurpx7koiti	18	12	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	18	13	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	18	14	now	now	RB
work_fklnz2kdpjhwfjysurpx7koiti	18	15	avail-	avail-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	18	16	able	able	JJ
work_fklnz2kdpjhwfjysurpx7koiti	18	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	19	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	19	2	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	19	3	lexico	lexico	JJ
work_fklnz2kdpjhwfjysurpx7koiti	19	4	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	19	5	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	19	6	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	7	extracted	extract	VBN
work_fklnz2kdpjhwfjysurpx7koiti	19	8	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	19	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	19	10	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	11	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	19	12	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	13	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	14	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	19	15	conjunction	conjunction	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	16	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	19	17	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	19	18	al-	al-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	19	19	gorithm	gorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	19	20	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	19	21	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	19	22	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	19	23	Mixed	Mixed	NNP
work_fklnz2kdpjhwfjysurpx7koiti	19	24	Integer	Integer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	19	25	Linear	Linear	NNP
work_fklnz2kdpjhwfjysurpx7koiti	19	26	Program	Program	NNP
work_fklnz2kdpjhwfjysurpx7koiti	19	27	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	19	28	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	19	29	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	19	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	20	1	Linguistic	linguistic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	2	analyses	analysis	NNS
work_fklnz2kdpjhwfjysurpx7koiti	20	3	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	20	4	identified	identify	VBN
work_fklnz2kdpjhwfjysurpx7koiti	20	5	phrases	phrase	NNS
work_fklnz2kdpjhwfjysurpx7koiti	20	6	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	7	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	8	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	9	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	20	10	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	20	11	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	12	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	20	13	hot	hot	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	20	15	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	20	16	scorch-	scorch-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	17	ing	ing	NNP
work_fklnz2kdpjhwfjysurpx7koiti	20	18	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	19	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	20	20	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	20	21	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	20	22	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	23	sources	source	NNS
work_fklnz2kdpjhwfjysurpx7koiti	20	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	25	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	20	26	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	27	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	20	28	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	20	29	intensities	intensity	NNS
work_fklnz2kdpjhwfjysurpx7koiti	20	30	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	20	31	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	20	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	21	1	However	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	21	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	21	3	pure	pure	JJ
work_fklnz2kdpjhwfjysurpx7koiti	21	4	infor-	infor-	XX
work_fklnz2kdpjhwfjysurpx7koiti	21	5	mation	mation	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	6	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	7	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	21	8	often	often	RB
work_fklnz2kdpjhwfjysurpx7koiti	21	9	fail	fail	VBP
work_fklnz2kdpjhwfjysurpx7koiti	21	10	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	21	11	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	21	12	enough	enough	JJ
work_fklnz2kdpjhwfjysurpx7koiti	21	13	coverage	coverage	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	14	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	21	15	real	real	JJ
work_fklnz2kdpjhwfjysurpx7koiti	21	16	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	21	17	world	world	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	18	downstream	downstream	JJ
work_fklnz2kdpjhwfjysurpx7koiti	21	19	appli-	appli-	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	20	cations	cation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	21	21	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	21	22	Tandon	Tandon	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	23	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	21	24	de	de	FW
work_fklnz2kdpjhwfjysurpx7koiti	21	25	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	21	27	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	21	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	21	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	21	30	unless	unless	IN
work_fklnz2kdpjhwfjysurpx7koiti	21	31	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	21	32	form	form	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	33	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	21	34	advanced	advanced	JJ
work_fklnz2kdpjhwfjysurpx7koiti	21	35	inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	21	36	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	21	37	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	21	38	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	21	39	Snow	Snow	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	40	et	et	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	41	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	42	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	43	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	21	44	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	21	45	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	21	46	Suchanek	Suchanek	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	47	et	et	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	48	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	49	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	21	50	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	21	51	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	21	52	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	21	53	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	22	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	2	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	22	3	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	22	5	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	22	6	address	address	VBP
work_fklnz2kdpjhwfjysurpx7koiti	22	7	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	8	sparsity	sparsity	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	9	problem	problem	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	10	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	11	relying	rely	VBG
work_fklnz2kdpjhwfjysurpx7koiti	22	12	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	13	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	14	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	22	15	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	16	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	17	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	22	18	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	22	19	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	20	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	22	21	model	model	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	22	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	22	23	extends	extend	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	22	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	25	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	26	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	27	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	28	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	29	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	22	30	com-	com-	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	31	plete	plete	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	32	joint	joint	JJ
work_fklnz2kdpjhwfjysurpx7koiti	22	33	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	34	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	35	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	36	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	37	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	38	continuous	continuous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	22	39	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	40	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	22	41	while	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	42	maintaining	maintain	VBG
work_fklnz2kdpjhwfjysurpx7koiti	22	43	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	22	44	constraints	constraint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	45	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	22	46	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	47	transi-	transi-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	22	48	tivity	tivity	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	49	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	22	50	giving	give	VBG
work_fklnz2kdpjhwfjysurpx7koiti	22	51	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	22	52	weight	weight	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	53	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	54	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	22	55	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	56	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	57	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	58	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	59	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	22	60	higher	high	JJR
work_fklnz2kdpjhwfjysurpx7koiti	22	61	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	22	62	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	22	63	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	22	64	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	23	1	Instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	23	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	23	3	considering	consider	VBG
work_fklnz2kdpjhwfjysurpx7koiti	23	4	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	5	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	23	6	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	23	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	23	8	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	9	deci-	deci-	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	10	sion	sion	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	11	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	23	13	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	23	14	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	23	15	exploit	exploit	VBP
work_fklnz2kdpjhwfjysurpx7koiti	23	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	23	17	fact	fact	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	18	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	23	19	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	23	20	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	23	21	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	23	22	benefit	benefit	VB
work_fklnz2kdpjhwfjysurpx7koiti	23	23	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	23	24	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	23	25	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	23	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	23	27	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	24	1	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	24	2	how	how	WRB
work_fklnz2kdpjhwfjysurpx7koiti	24	3	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	24	4	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	24	5	relate	relate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	24	6	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	24	7	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	24	8	third	third	JJ
work_fklnz2kdpjhwfjysurpx7koiti	24	9	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	24	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	25	1	Previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	25	2	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	25	3	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	25	4	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	25	6	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	25	8	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	25	9	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	25	10	Schulam	Schulam	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	25	12	Fellbaum	Fellbaum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	25	14	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	25	15	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	25	16	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	17	et	et	FW
work_fklnz2kdpjhwfjysurpx7koiti	25	18	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	19	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	25	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	25	21	2012	2012	CD
work_fklnz2kdpjhwfjysurpx7koiti	25	22	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	25	23	has	have	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	25	24	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	25	25	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	25	26	lexico	lexico	JJ
work_fklnz2kdpjhwfjysurpx7koiti	25	27	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	25	28	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	25	29	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	25	30	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	25	31	or-	or-	IN
work_fklnz2kdpjhwfjysurpx7koiti	25	32	279	279	CD
work_fklnz2kdpjhwfjysurpx7koiti	25	33	der	der	NN
work_fklnz2kdpjhwfjysurpx7koiti	25	34	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	25	35	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	26	1	They	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	26	2	mainly	mainly	RB
work_fklnz2kdpjhwfjysurpx7koiti	26	3	evaluate	evaluate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	26	4	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	26	5	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	6	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	26	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	26	8	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	26	10	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	11	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	26	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	26	13	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	26	14	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	26	15	present	present	VB
work_fklnz2kdpjhwfjysurpx7koiti	26	16	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	26	17	par-	par-	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	18	titioning	titioning	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	19	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	20	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	26	21	attempts	attempt	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	26	22	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	26	23	form	form	VB
work_fklnz2kdpjhwfjysurpx7koiti	26	24	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	26	25	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	26	26	placing	place	VBG
work_fklnz2kdpjhwfjysurpx7koiti	26	27	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	26	28	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	29	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	26	30	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	26	31	left	left	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	32	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	26	33	right	right	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	34	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	26	35	pivot	pivot	NN
work_fklnz2kdpjhwfjysurpx7koiti	26	36	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	26	37	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	27	1	Unfortunately	unfortunately	RB
work_fklnz2kdpjhwfjysurpx7koiti	27	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	27	3	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	27	4	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	5	often	often	RB
work_fklnz2kdpjhwfjysurpx7koiti	27	6	fails	fail	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	27	7	be-	be-	XX
work_fklnz2kdpjhwfjysurpx7koiti	27	8	cause	cause	VB
work_fklnz2kdpjhwfjysurpx7koiti	27	9	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	27	10	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	27	11	lack	lack	VBP
work_fklnz2kdpjhwfjysurpx7koiti	27	12	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	13	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	27	14	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	27	15	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	16	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	27	17	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	27	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	27	19	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	27	21	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	27	22	explained	explain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	27	23	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	27	24	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	27	25	detail	detail	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	26	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	27	27	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	27	28	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	27	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	28	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	28	2	contrast	contrast	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	28	4	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	28	5	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	28	6	jointly	jointly	RB
work_fklnz2kdpjhwfjysurpx7koiti	28	7	uses	use	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	28	8	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	9	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	28	10	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	28	11	relevant	relevant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	28	12	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	13	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	28	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	28	15	captures	capture	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	28	16	com-	com-	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	17	plex	plex	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	18	interactions	interaction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	28	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	28	20	inferences	inference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	28	21	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	28	22	produce	produce	VB
work_fklnz2kdpjhwfjysurpx7koiti	28	23	inten-	inten-	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	24	sity	sity	NN
work_fklnz2kdpjhwfjysurpx7koiti	28	25	scales	scale	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	28	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	29	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	29	2	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	29	3	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	29	4	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	29	5	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	29	6	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	29	7	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	29	8	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	29	9	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	29	10	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	29	11	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	29	12	there	there	EX
work_fklnz2kdpjhwfjysurpx7koiti	29	13	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	29	14	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	29	15	explicit	explicit	JJ
work_fklnz2kdpjhwfjysurpx7koiti	29	16	evi-	evi-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	29	17	dence	dence	NN
work_fklnz2kdpjhwfjysurpx7koiti	29	18	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	29	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	29	20	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	29	21	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	29	22	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	29	23	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	29	24	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	29	25	related	related	JJ
work_fklnz2kdpjhwfjysurpx7koiti	29	26	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	29	27	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	29	28	transitive	transitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	29	29	inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	29	30	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	29	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	30	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	30	2	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	30	3	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	30	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	30	5	flex-	flex-	NN
work_fklnz2kdpjhwfjysurpx7koiti	30	6	ible	ible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	30	7	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	30	8	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	30	9	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	30	10	incorporate	incorporate	VB
work_fklnz2kdpjhwfjysurpx7koiti	30	11	additional	additional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	30	12	synonymy	synonymy	NN
work_fklnz2kdpjhwfjysurpx7koiti	30	13	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	30	14	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	30	15	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	30	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	30	17	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	30	18	helps	help	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	30	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	30	20	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	30	21	find	find	VB
work_fklnz2kdpjhwfjysurpx7koiti	30	22	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	30	23	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	30	24	better	well	RBR
work_fklnz2kdpjhwfjysurpx7koiti	30	25	ranking	ranking	JJ
work_fklnz2kdpjhwfjysurpx7koiti	30	26	solution	solution	NN
work_fklnz2kdpjhwfjysurpx7koiti	30	27	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	30	28	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	31	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	31	2	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	31	3	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	31	4	extends	extend	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	31	5	easily	easily	RB
work_fklnz2kdpjhwfjysurpx7koiti	31	6	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	31	7	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	31	8	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	31	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	32	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	32	2	describe	describe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	32	3	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	32	4	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	32	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	32	6	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	32	7	multilingual	multilingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	32	8	extension	extension	NN
work_fklnz2kdpjhwfjysurpx7koiti	32	9	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	32	10	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	32	11	projection	projection	NN
work_fklnz2kdpjhwfjysurpx7koiti	32	12	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	32	13	cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	32	14	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	32	15	lingual	lingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	32	16	MILPs	milp	NNS
work_fklnz2kdpjhwfjysurpx7koiti	32	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	33	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	33	2	evaluate	evaluate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	33	3	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	33	4	predicted	predict	VBN
work_fklnz2kdpjhwfjysurpx7koiti	33	5	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	6	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	33	7	us-	us-	VBP
work_fklnz2kdpjhwfjysurpx7koiti	33	8	ing	e	VBG
work_fklnz2kdpjhwfjysurpx7koiti	33	9	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	33	10	pairwise	pairwise	VBP
work_fklnz2kdpjhwfjysurpx7koiti	33	11	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	12	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	33	14	rank-	rank-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	33	15	ing	ing	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	16	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	17	coefficients	coefficient	NNS
work_fklnz2kdpjhwfjysurpx7koiti	33	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	33	19	achieving	achieve	VBG
work_fklnz2kdpjhwfjysurpx7koiti	33	20	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	33	21	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	33	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	33	23	significantly	significantly	RB
work_fklnz2kdpjhwfjysurpx7koiti	33	24	better	well	JJR
work_fklnz2kdpjhwfjysurpx7koiti	33	25	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	33	26	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	33	27	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	33	28	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	29	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	33	30	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	33	31	&	&	CC
work_fklnz2kdpjhwfjysurpx7koiti	33	32	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	33	33	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	33	34	32	32	CD
work_fklnz2kdpjhwfjysurpx7koiti	33	35	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	36	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	33	37	error	error	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	38	reduc-	reduc-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	33	39	tion	tion	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	40	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	33	41	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	33	42	quite	quite	RB
work_fklnz2kdpjhwfjysurpx7koiti	33	43	close	close	RB
work_fklnz2kdpjhwfjysurpx7koiti	33	44	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	33	45	human	human	JJ
work_fklnz2kdpjhwfjysurpx7koiti	33	46	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	33	47	level	level	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	48	performance	performance	NN
work_fklnz2kdpjhwfjysurpx7koiti	33	49	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	34	1	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	34	2	Method	Method	NNP
work_fklnz2kdpjhwfjysurpx7koiti	34	3	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	34	4	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	34	5	section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	34	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	34	7	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	34	8	describe	describe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	34	9	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	34	10	step	step	NN
work_fklnz2kdpjhwfjysurpx7koiti	34	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	34	12	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	34	13	ap-	ap-	NN
work_fklnz2kdpjhwfjysurpx7koiti	34	14	proach	proach	NN
work_fklnz2kdpjhwfjysurpx7koiti	34	15	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	34	16	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	34	17	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	34	18	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	34	19	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	34	20	single	single	JJ
work_fklnz2kdpjhwfjysurpx7koiti	34	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	34	22	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	34	23	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	34	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	35	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	35	2	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	35	3	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	35	4	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	35	5	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	35	6	applied	apply	VBN
work_fklnz2kdpjhwfjysurpx7koiti	35	7	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	35	8	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	35	9	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	35	10	classes	class	NNS
work_fklnz2kdpjhwfjysurpx7koiti	35	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	35	12	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	35	13	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	35	14	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	35	15	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	35	16	English	English	NNP
work_fklnz2kdpjhwfjysurpx7koiti	35	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	36	1	2.1	2.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	36	2	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	36	3	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	36	4	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	36	5	Scoring	Scoring	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	6	Model	Model	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	7	2.1.1	2.1.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	36	8	Intensity	Intensity	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	9	Scales	Scales	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	10	Near	Near	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	11	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	36	12	synonyms	synonyms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	36	13	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	36	14	differ	differ	VB
work_fklnz2kdpjhwfjysurpx7koiti	36	15	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	36	16	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	36	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	36	18	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	37	1	joy	joy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	37	2	vs.	vs.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	37	3	euphoria	euphoria	NNP
work_fklnz2kdpjhwfjysurpx7koiti	37	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	37	5	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	37	6	drizzle	drizzle	NNP
work_fklnz2kdpjhwfjysurpx7koiti	37	7	vs.	vs.	IN
work_fklnz2kdpjhwfjysurpx7koiti	37	8	rain	rain	NN
work_fklnz2kdpjhwfjysurpx7koiti	37	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	38	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	38	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	38	3	particu-	particu-	VBG
work_fklnz2kdpjhwfjysurpx7koiti	38	4	larly	larly	RB
work_fklnz2kdpjhwfjysurpx7koiti	38	5	true	true	JJ
work_fklnz2kdpjhwfjysurpx7koiti	38	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	38	7	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	38	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	38	9	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	38	10	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	38	11	represent	represent	VB
work_fklnz2kdpjhwfjysurpx7koiti	38	12	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	38	13	degrees	degree	NNS
work_fklnz2kdpjhwfjysurpx7koiti	38	14	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	38	15	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	38	16	given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	38	17	quality	quality	NN
work_fklnz2kdpjhwfjysurpx7koiti	38	18	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	38	19	attribute	attribute	NN
work_fklnz2kdpjhwfjysurpx7koiti	38	20	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	38	21	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	38	22	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	38	23	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	38	24	age	age	NN
work_fklnz2kdpjhwfjysurpx7koiti	38	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	39	1	Many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	2	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	39	3	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	39	4	gradable	gradable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	39	6	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	39	7	al-	al-	RB
work_fklnz2kdpjhwfjysurpx7koiti	39	8	low	low	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	9	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	39	10	grading	grade	VBG
work_fklnz2kdpjhwfjysurpx7koiti	39	11	adverbial	adverbial	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	12	modifiers	modifier	NNS
work_fklnz2kdpjhwfjysurpx7koiti	39	13	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	39	14	express	express	VB
work_fklnz2kdpjhwfjysurpx7koiti	39	15	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	16	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	39	17	degrees	degree	NNS
work_fklnz2kdpjhwfjysurpx7koiti	39	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	39	19	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	39	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	39	21	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	39	22	house	house	NN
work_fklnz2kdpjhwfjysurpx7koiti	39	23	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	39	24	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	39	25	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	39	26	big	big	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	27	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	39	28	extremely	extremely	RB
work_fklnz2kdpjhwfjysurpx7koiti	39	29	big	big	JJ
work_fklnz2kdpjhwfjysurpx7koiti	39	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	40	1	Often	often	RB
work_fklnz2kdpjhwfjysurpx7koiti	40	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	3	however	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	40	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	5	completely	completely	RB
work_fklnz2kdpjhwfjysurpx7koiti	40	6	differ-	differ-	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	40	7	ent	ent	NN
work_fklnz2kdpjhwfjysurpx7koiti	40	8	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	40	9	refer	refer	VBP
work_fklnz2kdpjhwfjysurpx7koiti	40	10	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	40	11	varying	vary	VBG
work_fklnz2kdpjhwfjysurpx7koiti	40	12	degrees	degree	NNS
work_fklnz2kdpjhwfjysurpx7koiti	40	13	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	40	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	40	15	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	40	16	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	40	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	18	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	40	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	20	huge	huge	JJ
work_fklnz2kdpjhwfjysurpx7koiti	40	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	22	gigantic	gigantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	40	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	40	24	gargantuan	gargantuan	JJ
work_fklnz2kdpjhwfjysurpx7koiti	40	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	41	1	Even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	41	2	adjec-	adjec-	VBG
work_fklnz2kdpjhwfjysurpx7koiti	41	3	tives	tive	NNS
work_fklnz2kdpjhwfjysurpx7koiti	41	4	like	like	IN
work_fklnz2kdpjhwfjysurpx7koiti	41	5	enormous	enormous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	41	7	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	41	8	superb	superb	NNP
work_fklnz2kdpjhwfjysurpx7koiti	41	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	41	10	impossible	impossible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	11	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	41	12	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	41	13	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	41	14	considered	consider	VBN
work_fklnz2kdpjhwfjysurpx7koiti	41	15	non	non	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	16	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	17	gradable	gradable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	18	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	41	19	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	41	20	syntactic	syntactic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	21	perspec-	perspec-	NN
work_fklnz2kdpjhwfjysurpx7koiti	41	22	tive	tive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	41	23	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	41	24	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	41	25	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	41	26	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	41	27	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	41	28	such	such	PDT
work_fklnz2kdpjhwfjysurpx7koiti	41	29	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	41	30	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	41	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	42	1	Weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	42	2	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	42	3	Strong	Strong	NNP
work_fklnz2kdpjhwfjysurpx7koiti	42	4	Patterns	Patterns	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	42	5	Strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	42	6	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	42	7	Weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	42	8	Patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	42	9	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	43	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	43	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	43	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	43	4	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	43	5	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	43	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	44	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	44	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	45	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	45	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	45	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	45	4	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	45	5	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	46	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	47	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	47	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	47	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	47	4	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	47	5	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	47	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	48	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	48	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	49	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	49	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	49	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	49	4	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	49	5	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	49	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	50	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	51	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	51	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	51	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	51	4	although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	51	5	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	51	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	52	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	52	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	53	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	53	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	53	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	53	4	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	53	5	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	54	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	55	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	55	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	55	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	55	4	though	though	IN
work_fklnz2kdpjhwfjysurpx7koiti	55	5	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	55	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	56	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	56	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	57	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	57	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	57	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	57	4	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	57	5	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	57	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	58	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	59	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	59	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	59	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	59	4	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	59	5	and/or	and/or	CC
work_fklnz2kdpjhwfjysurpx7koiti	59	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	59	7	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	59	8	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	60	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	60	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	61	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	61	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	61	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	61	4	although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	61	5	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	61	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	62	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	63	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	63	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	63	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	63	4	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	63	5	and/or	and/or	CC
work_fklnz2kdpjhwfjysurpx7koiti	63	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	63	7	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	63	8	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	64	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	64	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	65	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	65	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	65	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	65	4	though	though	IN
work_fklnz2kdpjhwfjysurpx7koiti	65	5	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	65	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	66	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	66	2	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	66	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	67	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	67	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	68	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	69	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	69	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	69	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	69	4	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	69	5	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	69	6	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	70	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	70	2	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	70	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	71	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	71	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	72	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	72	2	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	72	3	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	72	4	Ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	72	5	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	72	6	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	72	7	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	72	8	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	72	9	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	72	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	73	1	Among	among	IN
work_fklnz2kdpjhwfjysurpx7koiti	73	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	73	3	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	73	4	represented	represent	VBN
work_fklnz2kdpjhwfjysurpx7koiti	73	5	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	73	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	73	7	regular	regular	JJ
work_fklnz2kdpjhwfjysurpx7koiti	73	8	expressions	expression	NNS
work_fklnz2kdpjhwfjysurpx7koiti	73	9	above	above	RB
work_fklnz2kdpjhwfjysurpx7koiti	73	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	73	11	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	73	12	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	73	13	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	73	14	those	those	DT
work_fklnz2kdpjhwfjysurpx7koiti	73	15	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	73	16	capture	capture	VBP
work_fklnz2kdpjhwfjysurpx7koiti	73	17	less	less	JJR
work_fklnz2kdpjhwfjysurpx7koiti	73	18	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	73	19	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	73	20	equal	equal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	73	21	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	73	22	five	five	CD
work_fklnz2kdpjhwfjysurpx7koiti	73	23	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	73	24	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	73	25	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	73	26	fit	fit	VB
work_fklnz2kdpjhwfjysurpx7koiti	73	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	73	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	73	29	Google	Google	NNP
work_fklnz2kdpjhwfjysurpx7koiti	73	30	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	73	31	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	73	32	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	73	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	73	34	see	see	VB
work_fklnz2kdpjhwfjysurpx7koiti	73	35	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	73	36	2.1.2	2.1.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	73	37	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	73	38	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	74	1	Articles	Articles	NNP
work_fklnz2kdpjhwfjysurpx7koiti	74	2	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	74	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	74	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	74	5	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	74	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	74	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	74	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	74	9	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	74	10	allowed	allow	VBN
work_fklnz2kdpjhwfjysurpx7koiti	74	11	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	74	12	appear	appear	VB
work_fklnz2kdpjhwfjysurpx7koiti	74	13	before	before	IN
work_fklnz2kdpjhwfjysurpx7koiti	74	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	74	15	wildcards	wildcard	NNS
work_fklnz2kdpjhwfjysurpx7koiti	74	16	wherever	wherever	WRB
work_fklnz2kdpjhwfjysurpx7koiti	74	17	possible	possible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	74	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	75	1	2.1.2	2.1.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	75	2	Intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	75	3	Patterns	Patterns	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	75	4	Linguistic	linguistic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	75	5	studies	study	NNS
work_fklnz2kdpjhwfjysurpx7koiti	75	6	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	75	7	found	find	VBN
work_fklnz2kdpjhwfjysurpx7koiti	75	8	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	75	9	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	75	10	like	like	IN
work_fklnz2kdpjhwfjysurpx7koiti	75	11	‘	'	''
work_fklnz2kdpjhwfjysurpx7koiti	75	12	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	76	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	76	2	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	76	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	76	4	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	76	5	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	76	6	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	77	1	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	77	2	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	77	3	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	77	4	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	77	5	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	77	6	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	77	7	reveal	reveal	VB
work_fklnz2kdpjhwfjysurpx7koiti	77	8	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	77	9	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	77	10	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	77	11	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	77	12	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	77	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	77	14	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	77	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	77	16	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	77	17	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	77	18	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	77	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	77	20	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	77	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	77	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	78	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	78	2	assume	assume	VBP
work_fklnz2kdpjhwfjysurpx7koiti	78	3	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	78	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	78	5	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	78	6	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	78	7	sets	set	NNS
work_fklnz2kdpjhwfjysurpx7koiti	78	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	78	9	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	78	10	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	78	11	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	78	12	allow	allow	VBP
work_fklnz2kdpjhwfjysurpx7koiti	78	13	us	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	78	14	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	78	15	infer	infer	VB
work_fklnz2kdpjhwfjysurpx7koiti	78	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	78	17	most	most	RBS
work_fklnz2kdpjhwfjysurpx7koiti	78	18	likely	likely	JJ
work_fklnz2kdpjhwfjysurpx7koiti	78	19	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	78	20	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	78	21	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	78	22	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	78	23	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	78	24	encoun-	encoun-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	78	25	tered	tere	VBD
work_fklnz2kdpjhwfjysurpx7koiti	78	26	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	78	27	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	78	28	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	78	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	79	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	2	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	3	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	4	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	79	6	Pws	Pws	NNP
work_fklnz2kdpjhwfjysurpx7koiti	79	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	79	8	contains	contain	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	79	9	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	79	10	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	79	11	reflect	reflect	VBP
work_fklnz2kdpjhwfjysurpx7koiti	79	12	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	13	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	14	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	79	15	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	16	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	17	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	79	18	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	19	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	79	21	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	79	23	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	24	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	25	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	26	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	79	27	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	79	28	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	79	29	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	30	sec-	sec-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	31	ond	ond	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	79	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	79	34	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	79	35	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	36	second	second	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	37	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	38	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	79	39	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	79	40	Psw	Psw	NNP
work_fklnz2kdpjhwfjysurpx7koiti	79	41	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	79	42	captures	capture	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	79	43	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	79	44	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	45	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	79	46	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	79	47	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	79	48	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	80	1	See	see	VB
work_fklnz2kdpjhwfjysurpx7koiti	80	2	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	3	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	80	4	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	80	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	80	6	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	80	7	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	8	terns	tern	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	80	9	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	80	10	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	80	11	used	use	VBD
work_fklnz2kdpjhwfjysurpx7koiti	80	12	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	80	13	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	80	14	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	80	16	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	80	17	see	see	VB
work_fklnz2kdpjhwfjysurpx7koiti	80	18	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	19	4.1	4.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	80	20	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	80	21	implementation	implementation	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	22	details	detail	NNS
work_fklnz2kdpjhwfjysurpx7koiti	80	23	regarding	regard	VBG
work_fklnz2kdpjhwfjysurpx7koiti	80	24	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	80	25	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	26	col-	col-	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	27	lection	lection	NN
work_fklnz2kdpjhwfjysurpx7koiti	80	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	80	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	81	1	Many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	81	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	81	3	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	81	4	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	81	5	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	81	6	apply	apply	VBP
work_fklnz2kdpjhwfjysurpx7koiti	81	7	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	81	8	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	81	9	parts	part	NNS
work_fklnz2kdpjhwfjysurpx7koiti	81	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	81	11	speech	speech	NN
work_fklnz2kdpjhwfjysurpx7koiti	81	12	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	81	13	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	82	1	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	82	2	drizzle	drizzle	NNP
work_fklnz2kdpjhwfjysurpx7koiti	82	3	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	82	4	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	82	5	rain	rain	NN
work_fklnz2kdpjhwfjysurpx7koiti	82	6	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	82	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	82	8	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	82	9	running	run	VBG
work_fklnz2kdpjhwfjysurpx7koiti	82	10	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	82	11	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	82	12	sprinting	sprint	VBG
work_fklnz2kdpjhwfjysurpx7koiti	82	13	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	82	14	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	82	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	82	16	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	82	17	significant	significant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	82	18	discrimination	discrimination	NN
work_fklnz2kdpjhwfjysurpx7koiti	82	19	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	82	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	82	21	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	82	22	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	82	23	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	82	24	right	right	JJ
work_fklnz2kdpjhwfjysurpx7koiti	82	25	direction	direction	NN
work_fklnz2kdpjhwfjysurpx7koiti	82	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	83	1	2.1.3	2.1.3	CD
work_fklnz2kdpjhwfjysurpx7koiti	83	2	Pairwise	Pairwise	NNP
work_fklnz2kdpjhwfjysurpx7koiti	83	3	Scores	Scores	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	83	4	Given	give	VBD
work_fklnz2kdpjhwfjysurpx7koiti	83	5	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	83	6	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	7	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	83	9	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	83	10	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	83	11	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	83	12	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	83	13	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	83	14	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	83	15	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	83	17	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	83	18	first	first	RB
work_fklnz2kdpjhwfjysurpx7koiti	83	19	collect	collect	VBP
work_fklnz2kdpjhwfjysurpx7koiti	83	20	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	21	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	83	22	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	83	23	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	24	or-	or-	XX
work_fklnz2kdpjhwfjysurpx7koiti	83	25	der	der	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	26	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	83	27	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	83	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	83	29	above	above	RB
work_fklnz2kdpjhwfjysurpx7koiti	83	30	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	83	31	mentioned	mention	VBN
work_fklnz2kdpjhwfjysurpx7koiti	83	32	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	33	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	83	34	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	83	35	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	83	36	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	83	37	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	83	38	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	39	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	83	40	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	41	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	42	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	83	43	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	84	1	Previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	2	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	3	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	84	4	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	5	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	6	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	84	7	limited	limited	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	84	9	sized	sized	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	10	raw	raw	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	11	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	12	corpora	corpora	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	13	revealed	reveal	VBD
work_fklnz2kdpjhwfjysurpx7koiti	84	14	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	84	15	cover-	cover-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	16	age	age	NN
work_fklnz2kdpjhwfjysurpx7koiti	84	17	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	84	18	often	often	RB
work_fklnz2kdpjhwfjysurpx7koiti	84	19	limited	limited	JJ
work_fklnz2kdpjhwfjysurpx7koiti	84	20	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	84	21	Hearst	Hearst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	84	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	84	23	1992	1992	CD
work_fklnz2kdpjhwfjysurpx7koiti	84	24	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	84	25	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	84	26	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	84	27	McKeown	McKeown	NNP
work_fklnz2kdpjhwfjysurpx7koiti	84	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	84	29	1993	1993	CD
work_fklnz2kdpjhwfjysurpx7koiti	84	30	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	84	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	85	1	Some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	85	2	studies	study	NNS
work_fklnz2kdpjhwfjysurpx7koiti	85	3	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	85	4	Chklovski	Chklovski	NNP
work_fklnz2kdpjhwfjysurpx7koiti	85	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	85	6	Pantel	Pantel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	85	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	85	8	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	85	9	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	85	10	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	85	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	85	12	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	85	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	85	14	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	85	15	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	85	16	used	use	VBD
work_fklnz2kdpjhwfjysurpx7koiti	85	17	hit	hit	VBD
work_fklnz2kdpjhwfjysurpx7koiti	85	18	counts	count	NNS
work_fklnz2kdpjhwfjysurpx7koiti	85	19	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	85	20	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	85	21	online	online	JJ
work_fklnz2kdpjhwfjysurpx7koiti	85	22	search	search	NN
work_fklnz2kdpjhwfjysurpx7koiti	85	23	engine	engine	NN
work_fklnz2kdpjhwfjysurpx7koiti	85	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	85	25	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	85	26	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	85	27	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	85	28	unstable	unstable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	85	29	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	85	30	irreproducible	irreproducible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	85	31	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	85	32	Kilgarriff	Kilgarriff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	85	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	85	34	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	85	35	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	85	36	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	86	1	To	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	86	2	avoid	avoid	VB
work_fklnz2kdpjhwfjysurpx7koiti	86	3	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	86	4	issues	issue	NNS
work_fklnz2kdpjhwfjysurpx7koiti	86	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	6	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	86	7	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	86	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	86	9	largest	large	JJS
work_fklnz2kdpjhwfjysurpx7koiti	86	10	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	11	280	280	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	12	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	13	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	15	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	16	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	17	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	18	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	20	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	23	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	25	minute	minute	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	86	27	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	29	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	86	30	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	31	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	32	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	33	24492.0	24492.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	34	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	35	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	36	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	37	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	38	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	39	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	40	248.0	248.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	41	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	42	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	43	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	44	minute	minute	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	45	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	46	97.0	97.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	47	good	good	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	48	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	49	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	86	50	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	51	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	52	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	53	1912.0	1912.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	54	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	55	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	86	56	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	57	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	58	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	59	89.0	89.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	60	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	61	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	62	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	63	minute	minute	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	64	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	65	41.0	41.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	66	good	good	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	67	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	68	though	though	IN
work_fklnz2kdpjhwfjysurpx7koiti	86	69	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	70	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	71	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	72	504.0	504.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	73	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	74	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	75	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	86	76	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	77	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	78	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	79	47.0	47.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	80	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	81	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	82	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	86	83	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	84	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	85	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	86	338.0	338.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	87	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	88	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	89	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	90	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	86	91	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	92	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	93	181.0	181.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	94	good	good	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	95	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	86	96	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	86	97	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	98	→	→	SYM
work_fklnz2kdpjhwfjysurpx7koiti	86	99	156.0	156.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	100	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	101	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	86	102	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	86	103	Some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	86	104	examples	example	NNS
work_fklnz2kdpjhwfjysurpx7koiti	86	105	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	86	106	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	86	107	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	108	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	86	109	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	110	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	111	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	86	112	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	113	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	86	114	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	86	115	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	86	116	phrases	phrase	NNS
work_fklnz2kdpjhwfjysurpx7koiti	86	117	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	86	118	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	86	119	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	86	120	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	87	1	static	static	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	2	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	87	4	counts	counts	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	87	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	87	7	Google	Google	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	8	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	87	10	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	87	11	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	12	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	87	13	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	87	15	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	87	17	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	87	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	87	20	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	87	21	contains	contain	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	87	22	English	English	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	23	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	24	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	87	25	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	87	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	87	27	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	87	28	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	87	29	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	30	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	87	31	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	87	33	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	87	34	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	87	35	observed	observe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	87	36	frequency	frequency	NN
work_fklnz2kdpjhwfjysurpx7koiti	87	37	counts	count	NNS
work_fklnz2kdpjhwfjysurpx7koiti	87	38	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	87	39	generated	generate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	87	40	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	87	41	nearly	nearly	RB
work_fklnz2kdpjhwfjysurpx7koiti	87	42	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	43	trillion	trillion	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	44	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	87	45	tokens	token	NNS
work_fklnz2kdpjhwfjysurpx7koiti	87	46	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	87	47	95	95	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	48	billion	billion	CD
work_fklnz2kdpjhwfjysurpx7koiti	87	49	sentences	sentence	NNS
work_fklnz2kdpjhwfjysurpx7koiti	87	50	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	88	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	88	2	consider	consider	VBP
work_fklnz2kdpjhwfjysurpx7koiti	88	3	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	88	4	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	88	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	88	6	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	88	7	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	88	8	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	88	9	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	88	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	88	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	88	12	in-	in-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	88	13	put	put	NN
work_fklnz2kdpjhwfjysurpx7koiti	88	14	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	88	15	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	88	16	turn	turn	NN
work_fklnz2kdpjhwfjysurpx7koiti	88	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	89	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	89	2	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	89	3	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	4	p	p	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	5	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	89	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	89	7	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	89	8	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	9	sets	set	NNS
work_fklnz2kdpjhwfjysurpx7koiti	89	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	89	11	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	89	13	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	14	Pws	Pws	NNP
work_fklnz2kdpjhwfjysurpx7koiti	89	15	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	89	16	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	17	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	89	18	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	19	Psw	Psw	NNP
work_fklnz2kdpjhwfjysurpx7koiti	89	20	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	89	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	89	22	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	89	23	in-	in-	RB
work_fklnz2kdpjhwfjysurpx7koiti	89	24	sert	sert	VBP
work_fklnz2kdpjhwfjysurpx7koiti	89	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	89	26	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	27	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	28	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	89	29	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	89	30	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	31	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	89	32	p(a1,a2	p(a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	89	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	89	34	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	89	35	get	get	VB
work_fklnz2kdpjhwfjysurpx7koiti	89	36	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	89	37	phrasal	phrasal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	38	query	query	NN
work_fklnz2kdpjhwfjysurpx7koiti	89	39	like	like	IN
work_fklnz2kdpjhwfjysurpx7koiti	89	40	“	"	``
work_fklnz2kdpjhwfjysurpx7koiti	89	41	big	big	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	42	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	89	43	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	89	44	huge	huge	JJ
work_fklnz2kdpjhwfjysurpx7koiti	89	45	”	"	''
work_fklnz2kdpjhwfjysurpx7koiti	89	46	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	90	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	90	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	90	3	done	do	VBN
work_fklnz2kdpjhwfjysurpx7koiti	90	4	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	90	5	replacing	replace	VBG
work_fklnz2kdpjhwfjysurpx7koiti	90	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	90	7	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	90	8	wildcards	wildcard	NNS
work_fklnz2kdpjhwfjysurpx7koiti	90	9	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	90	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	90	11	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	90	12	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	90	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	90	14	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	90	15	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	90	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	90	17	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	90	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	91	1	Finally	finally	RB
work_fklnz2kdpjhwfjysurpx7koiti	91	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	91	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	91	4	scan	scan	VBP
work_fklnz2kdpjhwfjysurpx7koiti	91	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	91	6	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	91	7	n-	n-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	91	8	grams	grams	NNP
work_fklnz2kdpjhwfjysurpx7koiti	91	9	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	91	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	91	11	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	91	12	batch	batch	NN
work_fklnz2kdpjhwfjysurpx7koiti	91	13	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	91	14	similar	similar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	91	15	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	91	16	Bansal	Bansal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	91	17	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	91	18	Klein	Klein	NNP
work_fklnz2kdpjhwfjysurpx7koiti	91	19	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	91	20	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	91	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	91	22	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	91	23	collect	collect	VB
work_fklnz2kdpjhwfjysurpx7koiti	91	24	frequencies	frequency	NNS
work_fklnz2kdpjhwfjysurpx7koiti	91	25	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	91	26	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	91	27	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	91	28	phrase	phrase	NN
work_fklnz2kdpjhwfjysurpx7koiti	91	29	queries	query	NNS
work_fklnz2kdpjhwfjysurpx7koiti	91	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	92	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	2	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	92	3	depicts	depict	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	92	4	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	92	5	examples	example	NNS
work_fklnz2kdpjhwfjysurpx7koiti	92	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	92	7	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	92	8	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	92	10	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	92	11	phrase	phrase	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	12	queries	query	NNS
work_fklnz2kdpjhwfjysurpx7koiti	92	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	92	14	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	92	15	fre-	fre-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	92	16	quencies	quencie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	92	17	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	92	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	92	19	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	20	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	92	21	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	22	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	92	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	93	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	93	2	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	93	3	collect	collect	VBP
work_fklnz2kdpjhwfjysurpx7koiti	93	4	frequencies	frequency	NNS
work_fklnz2kdpjhwfjysurpx7koiti	93	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	93	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	93	7	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	93	8	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	93	9	unigrams	unigram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	93	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	93	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	93	12	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	93	13	terns	tern	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	93	14	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	93	15	normalization	normalization	NN
work_fklnz2kdpjhwfjysurpx7koiti	93	16	purposes	purpose	NNS
work_fklnz2kdpjhwfjysurpx7koiti	93	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	94	1	Given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	94	2	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	94	3	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	4	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	5	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	6	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	7	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	8	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	94	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	94	10	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	11	count	count	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	12	function	function	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	13	cnt	cnt	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	94	15	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	94	16	define	define	VBP
work_fklnz2kdpjhwfjysurpx7koiti	94	17	W1	W1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	18	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	94	19	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	94	20	P1	p1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	21	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	22	p1∈Pws	p1∈Pws	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	23	cnt(p1(a1,a2	cnt(p1(a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	24	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	26	S1	s1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	27	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	94	28	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	94	29	P2	p2	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	30	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	31	p2∈Psw	p2∈psw	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	32	cnt(p2(a1,a2	cnt(p2(a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	34	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	35	W2	W2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	36	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	94	37	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	94	38	P1	p1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	39	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	40	p1∈Pws	p1∈Pws	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	41	cnt(p1(a2,a1	cnt(p1(a2,a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	42	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	43	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	44	S2	s2	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	45	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	94	46	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	94	47	P2	p2	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	48	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	49	p2∈Psw	p2∈psw	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	50	cnt(p2(a2,a1	cnt(p2(a2,a1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	51	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	52	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	53	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	54	1	1	LS
work_fklnz2kdpjhwfjysurpx7koiti	94	55	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	56	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	94	57	P1	p1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	58	=	=	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	59	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	60	p1∈Pws	p1∈Pws	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	61	cnt(p1	cnt(p1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	62	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	63	P2	p2	FW
work_fklnz2kdpjhwfjysurpx7koiti	94	64	=	=	NNS
work_fklnz2kdpjhwfjysurpx7koiti	94	65	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	94	66	p2∈Psw	p2∈psw	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	67	cnt(p2	cnt(p2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	68	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	69	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	94	70	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	71	2	2	LS
work_fklnz2kdpjhwfjysurpx7koiti	94	72	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	73	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	94	74	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	94	75	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	94	76	final	final	JJ
work_fklnz2kdpjhwfjysurpx7koiti	94	77	overall	overall	JJ
work_fklnz2kdpjhwfjysurpx7koiti	94	78	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	94	79	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	94	80	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	94	81	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	82	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	94	83	score(a1,a2	score(a1,a2	CD
work_fklnz2kdpjhwfjysurpx7koiti	94	84	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	85	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	94	86	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	87	W1	W1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	88	−S1	−s1	FW
work_fklnz2kdpjhwfjysurpx7koiti	94	89	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	90	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	91	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	92	W2	W2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	93	−S2	−S2	.
work_fklnz2kdpjhwfjysurpx7koiti	94	94	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	95	cnt(a1	cnt(a1	NN
work_fklnz2kdpjhwfjysurpx7koiti	94	96	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	97	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	94	98	cnt(a2	cnt(a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	94	99	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	94	100	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	95	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	95	2	3	3	LS
work_fklnz2kdpjhwfjysurpx7koiti	95	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	95	4	Here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	95	5	W1	W1	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	95	6	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	95	7	S1	S1	NNS
work_fklnz2kdpjhwfjysurpx7koiti	95	8	represent	represent	VBP
work_fklnz2kdpjhwfjysurpx7koiti	95	9	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	95	10	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	95	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	95	12	a1	a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	95	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	95	14	a2	a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	95	15	being	be	VBG
work_fklnz2kdpjhwfjysurpx7koiti	95	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	95	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	95	18	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	95	19	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	95	20	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	95	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	95	22	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	95	23	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	95	24	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	95	25	relation	relation	NN
work_fklnz2kdpjhwfjysurpx7koiti	95	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	95	27	respectively	respectively	RB
work_fklnz2kdpjhwfjysurpx7koiti	95	28	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	96	1	W2	W2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	96	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	96	3	S2	S2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	96	4	fit	fit	VBP
work_fklnz2kdpjhwfjysurpx7koiti	96	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	96	6	reverse	reverse	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	7	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	96	8	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	96	9	a2,a1	a2,a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	96	10	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	96	11	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	96	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	96	13	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	96	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	96	15	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	96	16	represent	represent	VB
work_fklnz2kdpjhwfjysurpx7koiti	96	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	96	18	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	19	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	96	20	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	96	22	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	23	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	96	24	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	25	relations	relation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	96	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	96	27	respec-	respec-	VBG
work_fklnz2kdpjhwfjysurpx7koiti	96	28	tively	tively	RB
work_fklnz2kdpjhwfjysurpx7koiti	96	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	96	30	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	96	31	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	96	32	opposite	opposite	JJ
work_fklnz2kdpjhwfjysurpx7koiti	96	33	direction	direction	NN
work_fklnz2kdpjhwfjysurpx7koiti	96	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	97	1	Hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	97	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	97	3	overall	overall	RB
work_fklnz2kdpjhwfjysurpx7koiti	97	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	97	5	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	6	W1	W1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	97	7	−S1	−s1	FW
work_fklnz2kdpjhwfjysurpx7koiti	97	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	9	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	97	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	11	W2	W2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	97	12	−S2	−S2	.
work_fklnz2kdpjhwfjysurpx7koiti	97	13	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	14	represents	represent	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	97	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	97	16	total	total	JJ
work_fklnz2kdpjhwfjysurpx7koiti	97	17	weak-	weak-	NN
work_fklnz2kdpjhwfjysurpx7koiti	97	18	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	97	19	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	97	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	97	21	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	97	22	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	97	23	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	24	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	97	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	97	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	97	27	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	98	1	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	98	2	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	98	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	98	4	a1	a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	98	5	being	be	VBG
work_fklnz2kdpjhwfjysurpx7koiti	98	6	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	98	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	98	8	left	left	NN
work_fklnz2kdpjhwfjysurpx7koiti	98	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	98	10	a2	a2	NN
work_fklnz2kdpjhwfjysurpx7koiti	98	11	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	98	12	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	98	13	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	98	14	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	98	15	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	98	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	98	17	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	98	18	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	98	19	score(a1,a2	score(a1,a2	CD
work_fklnz2kdpjhwfjysurpx7koiti	98	20	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	98	21	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	98	22	−score(a2,a1	−score(a2,a1	UH
work_fklnz2kdpjhwfjysurpx7koiti	98	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	98	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	99	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	2	raw	raw	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	3	frequencies	frequency	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	6	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	7	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	99	8	divided	divide	VBN
work_fklnz2kdpjhwfjysurpx7koiti	99	9	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	10	counts	count	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	13	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	99	15	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	16	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	17	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	18	unigram	unigram	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	19	counts	count	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	99	20	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	99	21	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	99	22	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	23	pointwise	pointwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	24	mutual	mutual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	25	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	99	27	PMI	PMI	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	99	29	style	style	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	30	normalization	normalization	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	31	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	99	32	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	99	33	avoid	avoid	VBP
work_fklnz2kdpjhwfjysurpx7koiti	99	34	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	35	bias	bias	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	36	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	37	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	38	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	39	due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	40	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	41	high	high	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	42	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	99	43	frequency	frequency	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	44	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	45	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	99	46	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	47	unigrams.2	unigrams.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	99	48	2.2	2.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	99	49	Global	Global	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	50	Ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	51	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	52	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	53	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	54	2.2.1	2.2.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	99	55	Objective	Objective	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	56	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	99	57	Constraints	Constraints	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	58	Given	Given	NNP
work_fklnz2kdpjhwfjysurpx7koiti	99	59	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	60	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	61	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	99	62	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	99	63	now	now	RB
work_fklnz2kdpjhwfjysurpx7koiti	99	64	aim	aim	VBP
work_fklnz2kdpjhwfjysurpx7koiti	99	65	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	66	producing	produce	VBG
work_fklnz2kdpjhwfjysurpx7koiti	99	67	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	68	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	69	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	70	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	71	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	72	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	73	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	74	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	99	75	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	99	76	much	much	RB
work_fklnz2kdpjhwfjysurpx7koiti	99	77	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	99	78	informative	informative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	79	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	99	80	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	99	81	original	original	JJ
work_fklnz2kdpjhwfjysurpx7koiti	99	82	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	99	83	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	99	84	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	100	1	Joint	joint	JJ
work_fklnz2kdpjhwfjysurpx7koiti	100	2	inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	3	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	4	multiple	multiple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	100	5	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	6	pairs	pairs	NNP
work_fklnz2kdpjhwfjysurpx7koiti	100	7	allows	allow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	100	8	us	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	100	9	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	100	10	ben-	ben-	VB
work_fklnz2kdpjhwfjysurpx7koiti	100	11	efit	efit	VB
work_fklnz2kdpjhwfjysurpx7koiti	100	12	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	13	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	100	14	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	15	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	100	16	Due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	17	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	100	19	sparsity	sparsity	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	21	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	100	22	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	23	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	100	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	100	25	determining	determine	VBG
work_fklnz2kdpjhwfjysurpx7koiti	100	26	how	how	WRB
work_fklnz2kdpjhwfjysurpx7koiti	100	27	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	100	28	adjec-	adjec-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	100	29	tives	tive	NNS
work_fklnz2kdpjhwfjysurpx7koiti	100	30	relate	relate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	100	31	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	100	32	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	100	33	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	100	34	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	100	35	sometimes	sometimes	RB
work_fklnz2kdpjhwfjysurpx7koiti	100	36	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	101	1	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	101	2	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	101	3	inferred	infer	VBN
work_fklnz2kdpjhwfjysurpx7koiti	101	4	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	101	5	observing	observe	VBG
work_fklnz2kdpjhwfjysurpx7koiti	101	6	how	how	WRB
work_fklnz2kdpjhwfjysurpx7koiti	101	7	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	101	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	101	9	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	101	10	relate	relate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	101	11	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	101	12	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	101	13	third	third	JJ
work_fklnz2kdpjhwfjysurpx7koiti	101	14	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	101	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	102	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	102	2	assume	assume	VBP
work_fklnz2kdpjhwfjysurpx7koiti	102	3	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	102	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	102	5	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	102	6	given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	102	7	N	n	CD
work_fklnz2kdpjhwfjysurpx7koiti	102	8	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	102	9	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	102	10	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	102	11	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	102	12	a1	a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	102	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	102	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	103	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	104	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	105	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	105	2	aN	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	105	3	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	105	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	105	5	wish	wish	VBP
work_fklnz2kdpjhwfjysurpx7koiti	105	6	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	105	7	place	place	VB
work_fklnz2kdpjhwfjysurpx7koiti	105	8	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	105	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	105	10	linear	linear	JJ
work_fklnz2kdpjhwfjysurpx7koiti	105	11	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	105	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	105	13	say	say	VBP
work_fklnz2kdpjhwfjysurpx7koiti	105	14	[	[	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	105	15	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	105	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	105	17	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	105	18	]	]	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	105	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	106	1	Thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	106	2	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	106	3	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	106	4	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	106	5	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	106	6	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	106	7	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	106	8	assigned	assign	VBN
work_fklnz2kdpjhwfjysurpx7koiti	106	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	106	10	position	position	NN
work_fklnz2kdpjhwfjysurpx7koiti	106	11	xi	xi	PRP
work_fklnz2kdpjhwfjysurpx7koiti	106	12	∈	∈	IN
work_fklnz2kdpjhwfjysurpx7koiti	106	13	[	[	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	106	14	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	106	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	106	16	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	106	17	]	]	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	106	18	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	106	19	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	106	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	106	21	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	106	22	weak-	weak-	NN
work_fklnz2kdpjhwfjysurpx7koiti	106	23	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	106	24	weights	weights	IN
work_fklnz2kdpjhwfjysurpx7koiti	106	25	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	106	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	106	27	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	106	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	106	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	107	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	107	2	positive	positive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	107	3	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	107	4	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	107	5	2In	2in	CD
work_fklnz2kdpjhwfjysurpx7koiti	107	6	preliminary	preliminary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	107	7	experiments	experiment	NNS
work_fklnz2kdpjhwfjysurpx7koiti	107	8	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	107	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	107	10	development	development	NN
work_fklnz2kdpjhwfjysurpx7koiti	107	11	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	107	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	107	13	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	107	14	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	107	15	evaluated	evaluate	VBD
work_fklnz2kdpjhwfjysurpx7koiti	107	16	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	107	17	intuitive	intuitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	107	18	forms	form	NNS
work_fklnz2kdpjhwfjysurpx7koiti	107	19	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	107	20	normalization	normalization	NN
work_fklnz2kdpjhwfjysurpx7koiti	107	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	108	1	281	281	CD
work_fklnz2kdpjhwfjysurpx7koiti	108	2	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	108	3	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	108	4	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	108	5	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	108	6	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	108	7	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	108	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	108	9	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	108	10	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	108	11	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	108	12	contain	contain	VB
work_fklnz2kdpjhwfjysurpx7koiti	108	13	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	108	14	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	108	15	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	108	16	cycles	cycle	NNS
work_fklnz2kdpjhwfjysurpx7koiti	108	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	108	18	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	109	1	due	due	JJ
work_fklnz2kdpjhwfjysurpx7koiti	109	2	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	109	3	noisy	noisy	JJ
work_fklnz2kdpjhwfjysurpx7koiti	109	4	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	109	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	109	6	so	so	CC
work_fklnz2kdpjhwfjysurpx7koiti	109	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	109	8	final	final	JJ
work_fklnz2kdpjhwfjysurpx7koiti	109	9	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	109	10	will	will	MD
work_fklnz2kdpjhwfjysurpx7koiti	109	11	have	have	VB
work_fklnz2kdpjhwfjysurpx7koiti	109	12	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	109	13	choose	choose	VB
work_fklnz2kdpjhwfjysurpx7koiti	109	14	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	109	15	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	109	16	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	109	17	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	109	18	honor	honor	VB
work_fklnz2kdpjhwfjysurpx7koiti	109	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	109	20	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	109	21	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	109	22	remove	remove	VB
work_fklnz2kdpjhwfjysurpx7koiti	109	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	110	1	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	110	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	110	3	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	110	4	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	110	5	means	mean	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	110	6	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	110	7	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	110	8	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	110	9	supposedly	supposedly	RB
work_fklnz2kdpjhwfjysurpx7koiti	110	10	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	110	11	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	110	12	aj	aj	PRP
work_fklnz2kdpjhwfjysurpx7koiti	110	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	110	14	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	110	15	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	110	16	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	110	17	like	like	VB
work_fklnz2kdpjhwfjysurpx7koiti	110	18	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	110	19	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	110	20	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	110	21	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	110	22	xj	xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	110	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	111	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	111	2	negative	negative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	111	3	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	111	4	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	111	5	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	111	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	111	7	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	111	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	111	9	means	mean	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	111	10	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	111	11	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	111	12	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	111	13	assumed	assume	VBN
work_fklnz2kdpjhwfjysurpx7koiti	111	14	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	111	15	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	111	16	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	111	17	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	111	18	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	111	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	111	20	so	so	RB
work_fklnz2kdpjhwfjysurpx7koiti	111	21	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	111	22	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	111	23	want	want	VB
work_fklnz2kdpjhwfjysurpx7koiti	111	24	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	111	25	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	111	26	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	111	27	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	111	28	xj	xj	XX
work_fklnz2kdpjhwfjysurpx7koiti	111	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	112	1	Therefore	therefore	RB
work_fklnz2kdpjhwfjysurpx7koiti	112	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	112	3	intuitively	intuitively	RB
work_fklnz2kdpjhwfjysurpx7koiti	112	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	112	5	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	112	6	goal	goal	NN
work_fklnz2kdpjhwfjysurpx7koiti	112	7	corresponds	correspond	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	112	8	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	112	9	maximizing	maximize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	112	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	112	11	objective	objective	NN
work_fklnz2kdpjhwfjysurpx7koiti	112	12	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	112	13	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	112	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	112	15	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	16	sgn(xj	sgn(xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	17	−xi	−xi	CD
work_fklnz2kdpjhwfjysurpx7koiti	112	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	19	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	112	20	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	112	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	112	22	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	24	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	25	4	4	LS
work_fklnz2kdpjhwfjysurpx7koiti	112	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	27	Note	note	VB
work_fklnz2kdpjhwfjysurpx7koiti	112	28	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	112	29	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	112	30	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	112	31	important	important	JJ
work_fklnz2kdpjhwfjysurpx7koiti	112	32	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	112	33	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	112	34	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	112	35	signum	signum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	36	func-	func-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	112	37	tion	tion	NN
work_fklnz2kdpjhwfjysurpx7koiti	112	38	sgn	sgn	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	39	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	40	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	112	41	here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	112	42	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	112	43	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	112	44	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	112	45	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	112	46	care	care	VBP
work_fklnz2kdpjhwfjysurpx7koiti	112	47	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	112	48	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	112	49	rel-	rel-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	50	ative	ative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	112	51	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	112	52	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	112	53	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	54	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	112	55	xj	xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	112	56	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	113	1	Maximizing	maximize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	113	2	∑	∑	CD
work_fklnz2kdpjhwfjysurpx7koiti	113	3	ij(xj	ij(xj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	113	4	−xi	−xi	CD
work_fklnz2kdpjhwfjysurpx7koiti	113	5	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	113	6	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	113	7	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	113	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	113	9	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	113	10	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	113	11	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	113	12	lead	lead	VB
work_fklnz2kdpjhwfjysurpx7koiti	113	13	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	113	14	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	113	15	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	113	16	being	be	VBG
work_fklnz2kdpjhwfjysurpx7koiti	113	17	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	113	18	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	113	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	113	20	edges	edge	NNS
work_fklnz2kdpjhwfjysurpx7koiti	113	21	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	113	22	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	113	23	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	113	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	113	25	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	113	26	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	113	27	highest	high	JJS
work_fklnz2kdpjhwfjysurpx7koiti	113	28	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	113	29	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	113	30	dominate	dominate	VB
work_fklnz2kdpjhwfjysurpx7koiti	113	31	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	113	32	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	113	33	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	113	34	ones	one	NNS
work_fklnz2kdpjhwfjysurpx7koiti	113	35	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	114	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	114	2	do	do	VBP
work_fklnz2kdpjhwfjysurpx7koiti	114	3	include	include	VB
work_fklnz2kdpjhwfjysurpx7koiti	114	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	114	5	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	114	6	magnitudes	magnitude	NNS
work_fklnz2kdpjhwfjysurpx7koiti	114	7	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	114	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	114	9	objective	objective	NN
work_fklnz2kdpjhwfjysurpx7koiti	114	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	114	11	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	114	12	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	114	13	help	help	VBP
work_fklnz2kdpjhwfjysurpx7koiti	114	14	resolve	resolve	VB
work_fklnz2kdpjhwfjysurpx7koiti	114	15	contradictions	contradiction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	114	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	114	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	114	18	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	114	19	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	114	20	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	114	21	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	114	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	114	23	see	see	VB
work_fklnz2kdpjhwfjysurpx7koiti	114	24	Figure	Figure	NNP
work_fklnz2kdpjhwfjysurpx7koiti	114	25	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	114	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	114	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	115	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	115	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	115	3	discussed	discuss	VBN
work_fklnz2kdpjhwfjysurpx7koiti	115	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	115	5	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	115	6	de-	de-	NN
work_fklnz2kdpjhwfjysurpx7koiti	115	7	tail	tail	NN
work_fklnz2kdpjhwfjysurpx7koiti	115	8	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	115	9	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	115	10	2.2.2	2.2.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	115	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	116	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	116	2	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	116	3	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	116	4	maximize	maximize	VB
work_fklnz2kdpjhwfjysurpx7koiti	116	5	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	116	6	non	non	JJ
work_fklnz2kdpjhwfjysurpx7koiti	116	7	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	116	8	differentiable	differentiable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	116	9	ob-	ob-	XX
work_fklnz2kdpjhwfjysurpx7koiti	116	10	jective	jective	NN
work_fklnz2kdpjhwfjysurpx7koiti	116	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	116	12	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	116	13	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	116	14	Mixed	Mixed	NNP
work_fklnz2kdpjhwfjysurpx7koiti	116	15	Integer	Integer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	116	16	Linear	Linear	NNP
work_fklnz2kdpjhwfjysurpx7koiti	116	17	Programming	Programming	NNP
work_fklnz2kdpjhwfjysurpx7koiti	116	18	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	116	19	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	116	20	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	116	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	116	22	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	116	23	variant	variant	NN
work_fklnz2kdpjhwfjysurpx7koiti	116	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	116	25	linear	linear	JJ
work_fklnz2kdpjhwfjysurpx7koiti	116	26	programming	programming	NN
work_fklnz2kdpjhwfjysurpx7koiti	116	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	116	28	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	116	29	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	116	30	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	116	31	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	116	32	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	116	33	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	116	34	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	116	35	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	116	36	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	116	37	constrained	constrain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	116	38	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	116	39	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	116	40	integers	integer	NNS
work_fklnz2kdpjhwfjysurpx7koiti	116	41	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	117	1	Using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	117	2	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	117	3	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	117	4	formalization	formalization	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	117	6	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	117	7	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	117	8	find	find	VB
work_fklnz2kdpjhwfjysurpx7koiti	117	9	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	117	10	globally	globally	RB
work_fklnz2kdpjhwfjysurpx7koiti	117	11	optimal	optimal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	12	solution	solution	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	13	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	117	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	117	15	joint	joint	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	16	deci-	deci-	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	17	sion	sion	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	18	space	space	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	117	20	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	117	21	unlike	unlike	IN
work_fklnz2kdpjhwfjysurpx7koiti	117	22	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	23	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	117	25	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	117	26	jointly	jointly	RB
work_fklnz2kdpjhwfjysurpx7koiti	117	27	ex-	ex-	VBD
work_fklnz2kdpjhwfjysurpx7koiti	117	28	ploit	ploit	NNP
work_fklnz2kdpjhwfjysurpx7koiti	117	29	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	30	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	31	rather	rather	RB
work_fklnz2kdpjhwfjysurpx7koiti	117	32	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	117	33	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	117	34	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	35	local	local	JJ
work_fklnz2kdpjhwfjysurpx7koiti	117	36	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	117	37	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	117	38	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	117	39	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	117	40	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	118	1	To	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	118	2	encode	encode	VB
work_fklnz2kdpjhwfjysurpx7koiti	118	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	118	4	objective	objective	NN
work_fklnz2kdpjhwfjysurpx7koiti	118	5	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	118	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	118	7	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	118	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	118	9	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	118	10	need	need	VBP
work_fklnz2kdpjhwfjysurpx7koiti	118	11	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	118	12	introduce	introduce	VB
work_fklnz2kdpjhwfjysurpx7koiti	118	13	additional	additional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	118	14	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	118	15	dij	dij	VBN
work_fklnz2kdpjhwfjysurpx7koiti	118	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	118	17	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	118	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	118	19	sij	sij	VBP
work_fklnz2kdpjhwfjysurpx7koiti	118	20	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	118	21	capture	capture	VB
work_fklnz2kdpjhwfjysurpx7koiti	118	22	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	118	23	effect	effect	NN
work_fklnz2kdpjhwfjysurpx7koiti	118	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	118	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	118	26	signum	signum	JJ
work_fklnz2kdpjhwfjysurpx7koiti	118	27	function	function	NN
work_fklnz2kdpjhwfjysurpx7koiti	118	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	118	29	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	118	30	explained	explain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	118	31	below	below	RB
work_fklnz2kdpjhwfjysurpx7koiti	118	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	119	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	119	2	additionally	additionally	RB
work_fklnz2kdpjhwfjysurpx7koiti	119	3	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	119	4	enable	enable	VBP
work_fklnz2kdpjhwfjysurpx7koiti	119	5	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	119	6	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	119	7	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	119	8	make	make	VB
work_fklnz2kdpjhwfjysurpx7koiti	119	9	use	use	NN
work_fklnz2kdpjhwfjysurpx7koiti	119	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	119	11	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	119	12	external	external	JJ
work_fklnz2kdpjhwfjysurpx7koiti	119	13	equivalence	equivalence	NN
work_fklnz2kdpjhwfjysurpx7koiti	119	14	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	119	15	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	119	16	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	119	17	infor-	infor-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	119	18	mation	mation	NN
work_fklnz2kdpjhwfjysurpx7koiti	119	19	E	E	NNS
work_fklnz2kdpjhwfjysurpx7koiti	119	20	⊆	⊆	:
work_fklnz2kdpjhwfjysurpx7koiti	119	21	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	119	22	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	119	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	119	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	120	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	121	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	122	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	122	2	N}×{1	N}×{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	122	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	122	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	123	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	124	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	125	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	125	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	125	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	125	4	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	125	5	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	125	6	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	125	7	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	125	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	126	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	126	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	126	3	context	context	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	126	5	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	126	6	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	126	7	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	126	8	considered	consider	VBN
work_fklnz2kdpjhwfjysurpx7koiti	126	9	synonymous	synonymous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	126	10	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	126	11	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	126	12	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	126	13	close	close	JJ
work_fklnz2kdpjhwfjysurpx7koiti	126	14	enough	enough	RB
work_fklnz2kdpjhwfjysurpx7koiti	126	15	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	126	16	meaning	meaning	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	17	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	126	18	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	126	19	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	126	20	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	126	21	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	126	22	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	126	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	126	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	126	25	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	126	26	position	position	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	126	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	126	29	inten-	inten-	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	30	sity	sity	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	31	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	126	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	127	1	If	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	127	2	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	127	3	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	127	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	127	5	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	127	7	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	8	E	E	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	127	10	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	127	11	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	127	12	safely	safely	RB
work_fklnz2kdpjhwfjysurpx7koiti	127	13	assume	assume	VB
work_fklnz2kdpjhwfjysurpx7koiti	127	14	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	127	15	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	127	17	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	18	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	127	19	near	near	JJ
work_fklnz2kdpjhwfjysurpx7koiti	127	20	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	127	21	equivalent	equivalent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	127	22	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	127	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	127	24	so	so	IN
work_fklnz2kdpjhwfjysurpx7koiti	127	25	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	127	26	should	should	MD
work_fklnz2kdpjhwfjysurpx7koiti	127	27	encourage	encourage	VB
work_fklnz2kdpjhwfjysurpx7koiti	127	28	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	127	30	xj	xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	127	31	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	127	32	remain	remain	VB
work_fklnz2kdpjhwfjysurpx7koiti	127	33	close	close	JJ
work_fklnz2kdpjhwfjysurpx7koiti	127	34	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	127	35	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	127	36	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	127	37	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	128	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	128	2	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	128	4	defined	define	VBN
work_fklnz2kdpjhwfjysurpx7koiti	128	5	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	128	6	follows	follow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	128	7	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	128	8	maximize∑	maximize∑	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	9	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	10	i	i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	128	12	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	13	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	14	6∈E	6∈e	CD
work_fklnz2kdpjhwfjysurpx7koiti	128	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	16	wij	wij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	17	−sij	−sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	128	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	19	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	128	20	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	128	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	128	22	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	128	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	24	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	25	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	128	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	27	i	i	NN
work_fklnz2kdpjhwfjysurpx7koiti	128	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	128	29	j)∈E	j)∈E	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	30	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	31	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	128	32	+	+	SYM
work_fklnz2kdpjhwfjysurpx7koiti	128	33	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	128	34	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	128	35	C	c	NN
work_fklnz2kdpjhwfjysurpx7koiti	128	36	subject	subject	JJ
work_fklnz2kdpjhwfjysurpx7koiti	128	37	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	128	38	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	39	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	128	40	xj	xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	41	−xi	−xi	CD
work_fklnz2kdpjhwfjysurpx7koiti	128	42	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	43	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	128	44	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	45	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	128	46	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	128	47	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	129	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	130	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	131	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	131	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	131	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	131	4	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	131	5	−wijC	−wijC	NNP
work_fklnz2kdpjhwfjysurpx7koiti	131	6	≤	≤	NN
work_fklnz2kdpjhwfjysurpx7koiti	131	7	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	131	8	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	131	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	131	10	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	131	11	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	131	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	131	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	132	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	133	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	134	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	134	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	134	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	134	4	dij	dij	NN
work_fklnz2kdpjhwfjysurpx7koiti	134	5	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	134	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	134	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	134	8	−wij)C	−wij)C	NNS
work_fklnz2kdpjhwfjysurpx7koiti	134	9	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	134	10	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	134	11	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	134	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	134	13	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	134	14	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	134	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	134	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	135	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	136	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	137	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	137	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	137	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	137	4	dij	dij	NN
work_fklnz2kdpjhwfjysurpx7koiti	137	5	+	+	SYM
work_fklnz2kdpjhwfjysurpx7koiti	137	6	sijC	sijc	NN
work_fklnz2kdpjhwfjysurpx7koiti	137	7	≥	≥	NN
work_fklnz2kdpjhwfjysurpx7koiti	137	8	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	137	9	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	137	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	137	11	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	137	12	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	137	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	137	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	138	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	139	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	140	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	140	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	140	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	140	4	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	5	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	140	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	140	8	−sij)C	−sij)C	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	9	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	140	10	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	140	11	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	140	13	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	14	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	140	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	140	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	141	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	142	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	143	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	143	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	143	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	143	4	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	143	5	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	143	6	[	[	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	143	7	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	143	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	143	9	1	1	LS
work_fklnz2kdpjhwfjysurpx7koiti	143	10	]	]	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	143	11	∀i	∀i	VB
work_fklnz2kdpjhwfjysurpx7koiti	143	12	∈{1	∈{1	NNS
work_fklnz2kdpjhwfjysurpx7koiti	143	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	143	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	144	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	145	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	146	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	146	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	146	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	146	4	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	146	5	∈{0	∈{0	NN
work_fklnz2kdpjhwfjysurpx7koiti	146	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	146	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	146	8	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	146	9	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	146	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	146	11	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	146	12	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	146	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	146	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	147	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	148	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	149	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	149	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	149	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	149	4	sij	sij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	149	5	∈{0	∈{0	NN
work_fklnz2kdpjhwfjysurpx7koiti	149	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	149	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	149	8	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	149	9	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	149	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	149	11	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	149	12	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	149	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	149	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	150	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	151	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	152	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	152	2	N	n	LS
work_fklnz2kdpjhwfjysurpx7koiti	152	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	152	4	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	152	5	difference	difference	NN
work_fklnz2kdpjhwfjysurpx7koiti	152	6	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	152	7	dij	dij	VBP
work_fklnz2kdpjhwfjysurpx7koiti	152	8	simply	simply	RB
work_fklnz2kdpjhwfjysurpx7koiti	152	9	capture	capture	VB
work_fklnz2kdpjhwfjysurpx7koiti	152	10	differ-	differ-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	152	11	ences	ence	NNS
work_fklnz2kdpjhwfjysurpx7koiti	152	12	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	152	13	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	152	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	152	15	xj	xj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	152	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	153	1	C	C	NNP
work_fklnz2kdpjhwfjysurpx7koiti	153	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	153	3	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	153	4	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	153	5	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	153	6	constant	constant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	153	7	greater	great	JJR
work_fklnz2kdpjhwfjysurpx7koiti	153	8	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	153	9	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	153	10	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	153	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	153	12	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	153	13	|score(ai	|score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	153	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	153	15	aj)|	aj)|	NNP
work_fklnz2kdpjhwfjysurpx7koiti	153	16	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	153	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	153	18	exact	exact	JJ
work_fklnz2kdpjhwfjysurpx7koiti	153	19	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	153	20	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	153	21	irrelevant	irrelevant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	153	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	154	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	154	2	indicator	indicator	NN
work_fklnz2kdpjhwfjysurpx7koiti	154	3	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	154	4	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	154	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	154	6	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	154	7	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	154	8	jointly	jointly	RB
work_fklnz2kdpjhwfjysurpx7koiti	154	9	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	154	10	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	154	11	determine	determine	VB
work_fklnz2kdpjhwfjysurpx7koiti	154	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	154	13	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	154	14	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	154	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	154	16	signum	signum	JJ
work_fklnz2kdpjhwfjysurpx7koiti	154	17	function	function	NN
work_fklnz2kdpjhwfjysurpx7koiti	154	18	sgn(dij	sgn(dij	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	154	19	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	154	20	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	154	21	sgn(xj	sgn(xj	ADD
work_fklnz2kdpjhwfjysurpx7koiti	154	22	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	154	23	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	154	24	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	154	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	155	1	Variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	155	2	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	155	3	become	become	VBP
work_fklnz2kdpjhwfjysurpx7koiti	155	4	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	155	5	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	155	6	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	155	7	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	155	8	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	155	9	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	155	10	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	155	11	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	155	12	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	155	13	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	155	14	serve	serve	VB
work_fklnz2kdpjhwfjysurpx7koiti	155	15	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	155	16	indicator	indicator	NN
work_fklnz2kdpjhwfjysurpx7koiti	155	17	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	155	18	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	155	19	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	155	20	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	155	21	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	155	22	relationships	relationship	NNS
work_fklnz2kdpjhwfjysurpx7koiti	155	23	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	155	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	155	25	output	output	NN
work_fklnz2kdpjhwfjysurpx7koiti	155	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	156	1	Variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	156	2	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	156	3	become	become	VBP
work_fklnz2kdpjhwfjysurpx7koiti	156	4	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	156	5	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	156	6	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	156	7	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	156	8	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	156	9	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	156	10	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	156	11	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	156	12	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	156	13	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	156	14	serve	serve	VB
work_fklnz2kdpjhwfjysurpx7koiti	156	15	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	156	16	indicator	indicator	NN
work_fklnz2kdpjhwfjysurpx7koiti	156	17	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	156	18	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	156	19	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	156	20	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	156	21	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	156	22	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	156	23	relationship	relationship	NN
work_fklnz2kdpjhwfjysurpx7koiti	156	24	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	156	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	156	26	output	output	NN
work_fklnz2kdpjhwfjysurpx7koiti	156	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	157	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	157	2	ob-	ob-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	157	3	jective	jective	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	4	encourages	encourage	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	157	5	wij	wij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	6	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	157	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	157	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	157	9	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	157	11	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	157	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	13	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	157	14	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	157	15	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	157	16	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	17	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	157	18	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	157	19	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	157	20	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	157	22	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	24	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	157	25	0.3	0.3	LS
work_fklnz2kdpjhwfjysurpx7koiti	157	26	When	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	157	27	equiva-	equiva-	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	28	lence	lence	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	29	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	30	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	31	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	32	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	33	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	157	34	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	157	35	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	157	36	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	157	37	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	157	38	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	39	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	157	40	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	157	41	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	42	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	157	43	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	157	44	E	e	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	45	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	157	46	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	47	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	157	48	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	157	49	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	157	50	wij	wij	NN
work_fklnz2kdpjhwfjysurpx7koiti	157	51	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	157	52	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	157	53	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	157	54	encouraged	encourage	VBN
work_fklnz2kdpjhwfjysurpx7koiti	157	55	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	158	1	2.2.2	2.2.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	158	2	Discussion	discussion	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	3	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	158	4	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	158	5	uses	use	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	158	6	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	7	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	158	9	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	158	10	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	11	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	158	12	together	together	RB
work_fklnz2kdpjhwfjysurpx7koiti	158	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	158	14	assimilates	assimilate	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	158	15	all	all	PDT
work_fklnz2kdpjhwfjysurpx7koiti	158	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	158	17	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	158	18	via	via	IN
work_fklnz2kdpjhwfjysurpx7koiti	158	19	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	158	20	transitivity	transitivity	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	21	constraints	constraint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	158	22	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	158	23	determine	determine	VB
work_fklnz2kdpjhwfjysurpx7koiti	158	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	158	25	posi-	posi-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	158	26	tions	tion	NNS
work_fklnz2kdpjhwfjysurpx7koiti	158	27	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	158	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	158	29	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	30	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	158	31	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	158	32	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	158	33	continuous	continuous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	158	34	real	real	RB
work_fklnz2kdpjhwfjysurpx7koiti	158	35	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	158	36	valued	value	VBN
work_fklnz2kdpjhwfjysurpx7koiti	158	37	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	158	38	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	159	1	Hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	159	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	159	3	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	159	4	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	5	addresses	address	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	159	6	drawbacks	drawback	NNS
work_fklnz2kdpjhwfjysurpx7koiti	159	7	3In	3in	CD
work_fklnz2kdpjhwfjysurpx7koiti	159	8	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	9	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	159	10	avoid	avoid	VB
work_fklnz2kdpjhwfjysurpx7koiti	159	11	numeric	numeric	JJ
work_fklnz2kdpjhwfjysurpx7koiti	159	12	instability	instability	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	13	issues	issue	NNS
work_fklnz2kdpjhwfjysurpx7koiti	159	14	due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	15	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	16	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	159	17	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	159	18	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	159	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	159	20	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	159	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	159	22	values	value	NNS
work_fklnz2kdpjhwfjysurpx7koiti	159	23	after	after	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	24	frequency	frequency	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	25	normalization	normalization	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	159	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	28	practice	practice	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	29	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	159	30	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	159	31	found	find	VBN
work_fklnz2kdpjhwfjysurpx7koiti	159	32	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	159	33	necessary	necessary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	159	34	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	159	35	rescale	rescale	VB
work_fklnz2kdpjhwfjysurpx7koiti	159	36	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	159	37	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	38	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	159	39	fac-	fac-	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	40	tor	tor	NN
work_fklnz2kdpjhwfjysurpx7koiti	159	41	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	42	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	159	43	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	159	44	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	159	45	smallest	small	JJS
work_fklnz2kdpjhwfjysurpx7koiti	159	46	|score(ai	|score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	159	47	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	159	48	aj)|	aj)|	NNP
work_fklnz2kdpjhwfjysurpx7koiti	159	49	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	159	50	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	159	51	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	160	1	282	282	CD
work_fklnz2kdpjhwfjysurpx7koiti	160	2	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	3	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	160	4	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	160	5	Equivalence	equivalence	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	6	Information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	7	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	160	8	Knowing	know	VBG
work_fklnz2kdpjhwfjysurpx7koiti	160	9	that	that	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	10	am	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	12	a2	a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	160	13	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	14	synonyms	synonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	15	gives	give	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	160	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	17	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	160	18	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	19	indication	indication	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	21	where	where	WRB
work_fklnz2kdpjhwfjysurpx7koiti	160	22	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	160	23	place	place	VB
work_fklnz2kdpjhwfjysurpx7koiti	160	24	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	25	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	26	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	27	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	28	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	29	respect	respect	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	30	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	31	a1	a1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	160	32	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	33	a2	a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	160	34	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	35	a3	a3	NNP
work_fklnz2kdpjhwfjysurpx7koiti	160	36	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	37	local	local	JJ
work_fklnz2kdpjhwfjysurpx7koiti	160	38	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	160	39	divide	divide	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	40	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	160	41	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	160	42	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	160	43	conquer	conquer	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	44	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	160	45	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	46	where	where	WRB
work_fklnz2kdpjhwfjysurpx7koiti	160	47	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	160	48	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	49	scored	score	VBN
work_fklnz2kdpjhwfjysurpx7koiti	160	50	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	51	respect	respect	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	52	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	53	selected	select	VBN
work_fklnz2kdpjhwfjysurpx7koiti	160	54	pivot	pivot	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	55	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	160	56	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	57	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	160	58	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	160	59	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	160	60	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	160	61	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	62	lack	lack	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	63	pairwise	pairwise	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	64	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	65	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	66	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	67	pivots	pivot	NNS
work_fklnz2kdpjhwfjysurpx7koiti	160	68	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	160	69	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	160	70	properly	properly	RB
work_fklnz2kdpjhwfjysurpx7koiti	160	71	classified	classify	VBN
work_fklnz2kdpjhwfjysurpx7koiti	160	72	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	160	73	although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	74	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	160	75	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	160	76	have	have	VB
work_fklnz2kdpjhwfjysurpx7koiti	160	77	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	78	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	79	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	160	80	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	81	third	third	JJ
work_fklnz2kdpjhwfjysurpx7koiti	160	82	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	83	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	160	84	could	could	MD
work_fklnz2kdpjhwfjysurpx7koiti	160	85	help	help	VB
work_fklnz2kdpjhwfjysurpx7koiti	160	86	establish	establish	VB
work_fklnz2kdpjhwfjysurpx7koiti	160	87	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	160	88	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	160	89	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	161	1	Optional	optional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	161	2	synonymy	synonymy	NN
work_fklnz2kdpjhwfjysurpx7koiti	161	3	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	161	4	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	161	5	further	further	RB
work_fklnz2kdpjhwfjysurpx7koiti	161	6	help	help	VB
work_fklnz2kdpjhwfjysurpx7koiti	161	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	161	8	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	161	9	shown	show	VBN
work_fklnz2kdpjhwfjysurpx7koiti	161	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	161	11	Figure	Figure	NNP
work_fklnz2kdpjhwfjysurpx7koiti	161	12	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	161	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	162	1	Moreover	moreover	RB
work_fklnz2kdpjhwfjysurpx7koiti	162	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	162	3	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	162	4	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	162	5	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	162	6	gives	give	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	162	7	higher	high	JJR
work_fklnz2kdpjhwfjysurpx7koiti	162	8	weight	weight	NN
work_fklnz2kdpjhwfjysurpx7koiti	162	9	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	162	10	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	162	11	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	162	12	higher	high	JJR
work_fklnz2kdpjhwfjysurpx7koiti	162	13	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	162	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	162	15	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	162	16	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	162	17	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	162	18	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	162	19	breaking	break	VBG
work_fklnz2kdpjhwfjysurpx7koiti	162	20	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	162	21	constraint	constraint	NN
work_fklnz2kdpjhwfjysurpx7koiti	162	22	cycles	cycle	NNS
work_fklnz2kdpjhwfjysurpx7koiti	162	23	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	162	24	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	162	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	162	26	simple	simple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	162	27	example	example	NN
work_fklnz2kdpjhwfjysurpx7koiti	162	28	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	162	29	Figure	Figure	NNP
work_fklnz2kdpjhwfjysurpx7koiti	162	30	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	162	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	163	1	If	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	163	2	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	163	3	need	need	VBP
work_fklnz2kdpjhwfjysurpx7koiti	163	4	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	163	5	break	break	VB
work_fklnz2kdpjhwfjysurpx7koiti	163	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	163	7	con-	con-	NN
work_fklnz2kdpjhwfjysurpx7koiti	163	8	straint	straint	NN
work_fklnz2kdpjhwfjysurpx7koiti	163	9	violating	violate	VBG
work_fklnz2kdpjhwfjysurpx7koiti	163	10	triangle	triangle	NN
work_fklnz2kdpjhwfjysurpx7koiti	163	11	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	163	12	cycle	cycle	NN
work_fklnz2kdpjhwfjysurpx7koiti	163	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	163	14	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	163	15	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	163	16	have	have	VB
work_fklnz2kdpjhwfjysurpx7koiti	163	17	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	163	18	make	make	VB
work_fklnz2kdpjhwfjysurpx7koiti	163	19	arbitrary	arbitrary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	163	20	choices	choice	NNS
work_fklnz2kdpjhwfjysurpx7koiti	163	21	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	163	22	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	163	23	were	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	163	24	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	163	25	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	163	26	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	163	27	sgn(score(a	sgn(score(a	NN
work_fklnz2kdpjhwfjysurpx7koiti	163	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	163	29	b	b	LS
work_fklnz2kdpjhwfjysurpx7koiti	163	30	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	163	31	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	163	32	alone	alone	RB
work_fklnz2kdpjhwfjysurpx7koiti	163	33	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	164	1	Instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	164	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	164	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	164	4	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	164	5	choose	choose	VB
work_fklnz2kdpjhwfjysurpx7koiti	164	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	164	7	better	well	RBR
work_fklnz2kdpjhwfjysurpx7koiti	164	8	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	164	9	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	164	10	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	164	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	164	12	magnitude	magnitude	NN
work_fklnz2kdpjhwfjysurpx7koiti	164	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	164	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	164	15	pair-	pair-	NN
work_fklnz2kdpjhwfjysurpx7koiti	164	16	wise	wise	JJ
work_fklnz2kdpjhwfjysurpx7koiti	164	17	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	164	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	165	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	165	2	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	165	3	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	165	4	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	165	5	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	165	6	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	165	7	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	165	8	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	165	9	n’t	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	165	10	necessarily	necessarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	165	11	mean	mean	VB
work_fklnz2kdpjhwfjysurpx7koiti	165	12	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	165	13	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	165	14	should	should	MD
work_fklnz2kdpjhwfjysurpx7koiti	165	15	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	165	16	further	far	RBR
work_fklnz2kdpjhwfjysurpx7koiti	165	17	apart	apart	RB
work_fklnz2kdpjhwfjysurpx7koiti	165	18	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	165	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	165	20	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	165	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	166	1	It	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	166	2	means	mean	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	166	3	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	4	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	166	5	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	166	6	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	166	7	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	166	8	attested	attest	VBN
work_fklnz2kdpjhwfjysurpx7koiti	166	9	together	together	RB
work_fklnz2kdpjhwfjysurpx7koiti	166	10	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	166	12	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	166	13	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	14	respect	respect	NN
work_fklnz2kdpjhwfjysurpx7koiti	166	15	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	166	17	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	166	18	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	166	19	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	166	20	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	21	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	166	22	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	166	23	candi-	candi-	NN
work_fklnz2kdpjhwfjysurpx7koiti	166	24	date	date	NN
work_fklnz2kdpjhwfjysurpx7koiti	166	25	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	166	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	167	1	Therefore	therefore	RB
work_fklnz2kdpjhwfjysurpx7koiti	167	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	167	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	167	4	try	try	VBP
work_fklnz2kdpjhwfjysurpx7koiti	167	5	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	167	6	respect	respect	VB
work_fklnz2kdpjhwfjysurpx7koiti	167	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	167	8	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	167	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	167	10	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	167	11	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	167	12	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	167	13	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	167	14	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	167	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	167	16	final	final	JJ
work_fklnz2kdpjhwfjysurpx7koiti	167	17	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	167	18	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	167	19	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	167	20	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	167	21	breaking	break	VBG
work_fklnz2kdpjhwfjysurpx7koiti	167	22	constraint	constraint	NN
work_fklnz2kdpjhwfjysurpx7koiti	167	23	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	167	24	violating	violate	VBG
work_fklnz2kdpjhwfjysurpx7koiti	167	25	cycles	cycle	NNS
work_fklnz2kdpjhwfjysurpx7koiti	167	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	168	1	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	168	2	Related	Related	NNP
work_fklnz2kdpjhwfjysurpx7koiti	168	3	Work	Work	NNP
work_fklnz2kdpjhwfjysurpx7koiti	168	4	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	168	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	168	6	McKeown	McKeown	NNP
work_fklnz2kdpjhwfjysurpx7koiti	168	7	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	168	8	1993	1993	CD
work_fklnz2kdpjhwfjysurpx7koiti	168	9	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	168	10	presented	present	VBD
work_fklnz2kdpjhwfjysurpx7koiti	168	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	168	12	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	168	13	step	step	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	14	towards	towards	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	15	automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	168	16	identification	identification	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	17	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	18	ad-	ad-	CC
work_fklnz2kdpjhwfjysurpx7koiti	168	19	jective	jective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	168	20	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	168	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	168	22	thoroughly	thoroughly	RB
work_fklnz2kdpjhwfjysurpx7koiti	168	23	discussing	discuss	VBG
work_fklnz2kdpjhwfjysurpx7koiti	168	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	168	25	background	background	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	26	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	27	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	168	28	semantics	semantic	NNS
work_fklnz2kdpjhwfjysurpx7koiti	168	29	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	168	30	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	168	31	means	means	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	32	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	33	discovering	discover	VBG
work_fklnz2kdpjhwfjysurpx7koiti	168	34	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	168	35	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	36	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	168	37	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	168	38	belong	belong	VBP
work_fklnz2kdpjhwfjysurpx7koiti	168	39	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	40	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	168	41	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	168	42	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	43	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	168	44	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	168	45	providing	provide	VBG
work_fklnz2kdpjhwfjysurpx7koiti	168	46	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	168	47	way	way	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	48	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	49	creating	create	VBG
work_fklnz2kdpjhwfjysurpx7koiti	168	50	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	168	51	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	52	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	168	53	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	168	54	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	168	55	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	168	56	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	169	1	Inkpen	Inkpen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	169	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	169	3	Hirst	Hirst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	169	4	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	169	5	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	169	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	169	7	study	study	VB
work_fklnz2kdpjhwfjysurpx7koiti	169	8	near	near	JJ
work_fklnz2kdpjhwfjysurpx7koiti	169	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	169	10	synonyms	synonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	169	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	169	12	nuances	nuance	NNS
work_fklnz2kdpjhwfjysurpx7koiti	169	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	169	14	meaning	mean	VBG
work_fklnz2kdpjhwfjysurpx7koiti	169	15	differentiation	differentiation	NN
work_fklnz2kdpjhwfjysurpx7koiti	169	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	169	17	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	169	18	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	169	19	stylistic	stylistic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	169	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	169	21	attitudinal	attitudinal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	169	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	169	23	etc	etc	FW
work_fklnz2kdpjhwfjysurpx7koiti	169	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	169	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	169	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	170	1	They	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	170	2	attempt	attempt	VBP
work_fklnz2kdpjhwfjysurpx7koiti	170	3	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	170	4	automatically	automatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	170	5	ac-	ac-	RB
work_fklnz2kdpjhwfjysurpx7koiti	170	6	quire	quire	VB
work_fklnz2kdpjhwfjysurpx7koiti	170	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	170	8	knowledge	knowledge	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	9	base	base	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	170	11	near	near	IN
work_fklnz2kdpjhwfjysurpx7koiti	170	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	170	13	synonym	synonym	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	14	differences	difference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	170	15	via	via	IN
work_fklnz2kdpjhwfjysurpx7koiti	170	16	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	170	17	unsupervised	unsupervised	JJ
work_fklnz2kdpjhwfjysurpx7koiti	170	18	decision	decision	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	19	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	170	20	list	list	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	21	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	170	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	171	1	How-	How-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	171	2	ever	ever	RB
work_fklnz2kdpjhwfjysurpx7koiti	171	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	171	4	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	171	5	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	6	depends	depend	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	171	7	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	171	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	171	9	special	special	JJ
work_fklnz2kdpjhwfjysurpx7koiti	171	10	dictionary	dictionary	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	171	12	synonym	synonym	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	13	differences	difference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	171	14	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	171	15	learn	learn	VB
work_fklnz2kdpjhwfjysurpx7koiti	171	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	171	17	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	18	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	19	terns	tern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	171	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	171	21	while	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	171	22	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	171	23	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	171	24	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	171	25	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	171	26	raw	raw	JJ
work_fklnz2kdpjhwfjysurpx7koiti	171	27	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	28	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	171	29	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	30	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	171	31	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	172	1	Mohammad	Mohammad	NNP
work_fklnz2kdpjhwfjysurpx7koiti	172	2	et	et	NNP
work_fklnz2kdpjhwfjysurpx7koiti	172	3	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	172	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	173	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	173	2	2013	2013	CD
work_fklnz2kdpjhwfjysurpx7koiti	173	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	173	4	proposed	propose	VBD
work_fklnz2kdpjhwfjysurpx7koiti	173	5	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	173	6	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	173	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	173	8	identifying	identify	VBG
work_fklnz2kdpjhwfjysurpx7koiti	173	9	whether	whether	IN
work_fklnz2kdpjhwfjysurpx7koiti	173	10	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	173	11	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	173	12	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	173	13	antonymous	antonymous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	173	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	174	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	174	2	problem	problem	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	174	4	related	related	JJ
work_fklnz2kdpjhwfjysurpx7koiti	174	5	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	174	6	distinct	distinct	JJ
work_fklnz2kdpjhwfjysurpx7koiti	174	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	174	8	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	174	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	174	10	de-	de-	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	11	gree	gree	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	174	13	antonymy	antonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	174	14	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	174	15	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	174	16	necessarily	necessarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	174	17	determine	determine	VB
work_fklnz2kdpjhwfjysurpx7koiti	174	18	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	174	19	position	position	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	20	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	174	21	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	174	22	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	23	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	174	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	175	1	Antonyms	Antonyms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	175	2	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	175	3	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	175	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	175	5	little	little	JJ
work_fklnz2kdpjhwfjysurpx7koiti	175	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	175	7	big	big	JJ
work_fklnz2kdpjhwfjysurpx7koiti	175	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	175	9	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	175	10	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	175	11	necessarily	necessarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	175	12	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	175	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	175	14	extreme	extreme	JJ
work_fklnz2kdpjhwfjysurpx7koiti	175	15	ends	end	NNS
work_fklnz2kdpjhwfjysurpx7koiti	175	16	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	175	17	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	175	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	176	1	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	176	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	176	3	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	176	4	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	176	5	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	176	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	176	7	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	176	8	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	176	9	et	et	FW
work_fklnz2kdpjhwfjysurpx7koiti	176	10	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	176	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	177	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	177	2	2012	2012	CD
work_fklnz2kdpjhwfjysurpx7koiti	177	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	177	4	present	present	VB
work_fklnz2kdpjhwfjysurpx7koiti	177	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	177	6	most	most	RBS
work_fklnz2kdpjhwfjysurpx7koiti	177	7	closely	closely	RB
work_fklnz2kdpjhwfjysurpx7koiti	177	8	related	relate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	177	9	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	177	10	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	177	11	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	177	12	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	177	13	intensities	intensity	NNS
work_fklnz2kdpjhwfjysurpx7koiti	177	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	178	1	They	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	178	2	collect	collect	VBP
work_fklnz2kdpjhwfjysurpx7koiti	178	3	lexico-	lexico-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	178	4	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	178	5	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	178	6	via	via	IN
work_fklnz2kdpjhwfjysurpx7koiti	178	7	bootstrapping	bootstrappe	VBG
work_fklnz2kdpjhwfjysurpx7koiti	178	8	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	178	9	seed	seed	NN
work_fklnz2kdpjhwfjysurpx7koiti	178	10	adjec-	adjec-	NN
work_fklnz2kdpjhwfjysurpx7koiti	178	11	tive	tive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	178	12	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	178	13	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	178	14	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	178	15	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	178	16	intensities	intensity	NNS
work_fklnz2kdpjhwfjysurpx7koiti	178	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	178	18	albeit	albeit	IN
work_fklnz2kdpjhwfjysurpx7koiti	178	19	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	178	20	search	search	NN
work_fklnz2kdpjhwfjysurpx7koiti	178	21	engine	engine	NN
work_fklnz2kdpjhwfjysurpx7koiti	178	22	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	178	23	hits	hit	NNS
work_fklnz2kdpjhwfjysurpx7koiti	178	24	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	178	25	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	178	26	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	178	27	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	178	28	unstable	unstable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	178	29	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	178	30	prob-	prob-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	178	31	lematic	lematic	NNP
work_fklnz2kdpjhwfjysurpx7koiti	178	32	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	178	33	Kilgarriff	Kilgarriff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	178	34	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	178	35	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	178	36	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	178	37	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	179	1	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	179	2	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	179	3	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	179	5	primarily	primarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	179	6	evaluated	evaluate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	179	7	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	179	8	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	179	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	179	10	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	179	11	local	local	JJ
work_fklnz2kdpjhwfjysurpx7koiti	179	12	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	13	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	14	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	179	16	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	179	17	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	179	18	suggest	suggest	VBP
work_fklnz2kdpjhwfjysurpx7koiti	179	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	179	20	possibil-	possibil-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	179	21	ity	ity	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	22	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	179	23	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	179	24	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	179	25	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	179	26	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	179	27	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	28	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	179	29	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	179	30	pivot-	pivot-	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	31	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	179	32	partitioning	partition	VBG
work_fklnz2kdpjhwfjysurpx7koiti	179	33	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	179	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	180	1	Although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	180	2	intuitive	intuitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	180	3	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	180	4	theory	theory	NN
work_fklnz2kdpjhwfjysurpx7koiti	180	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	180	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	180	7	extracted	extracted	JJ
work_fklnz2kdpjhwfjysurpx7koiti	180	8	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	180	9	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	180	10	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	180	11	frequently	frequently	RB
work_fklnz2kdpjhwfjysurpx7koiti	180	12	too	too	RB
work_fklnz2kdpjhwfjysurpx7koiti	180	13	sparse	sparse	JJ
work_fklnz2kdpjhwfjysurpx7koiti	180	14	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	180	15	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	180	16	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	180	17	work	work	VB
work_fklnz2kdpjhwfjysurpx7koiti	180	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	181	1	Thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	181	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	181	3	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	181	4	adjec-	adjec-	NN
work_fklnz2kdpjhwfjysurpx7koiti	181	5	tives	tive	NNS
work_fklnz2kdpjhwfjysurpx7koiti	181	6	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	181	7	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	181	8	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	181	9	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	181	10	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	181	11	particular	particular	JJ
work_fklnz2kdpjhwfjysurpx7koiti	181	12	headword	headword	NN
work_fklnz2kdpjhwfjysurpx7koiti	181	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	182	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	182	2	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	182	3	experiments	experiment	NNS
work_fklnz2kdpjhwfjysurpx7koiti	182	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	182	5	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	182	6	reimplemented	reimplemente	VBD
work_fklnz2kdpjhwfjysurpx7koiti	182	7	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	182	8	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	182	9	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	182	10	show	show	VBP
work_fklnz2kdpjhwfjysurpx7koiti	182	11	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	182	12	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	182	13	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	182	14	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	182	15	improves	improve	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	182	16	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	182	17	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	182	18	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	182	19	allowing	allow	VBG
work_fklnz2kdpjhwfjysurpx7koiti	182	20	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	182	21	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	182	22	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	182	23	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	182	24	benefit	benefit	VB
work_fklnz2kdpjhwfjysurpx7koiti	182	25	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	182	26	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	182	27	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	182	28	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	182	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	183	1	Schulam	Schulam	NNP
work_fklnz2kdpjhwfjysurpx7koiti	183	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	183	3	Fell-	Fell-	VBN
work_fklnz2kdpjhwfjysurpx7koiti	183	4	baum	baum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	183	5	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	183	6	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	183	7	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	183	8	apply	apply	VBP
work_fklnz2kdpjhwfjysurpx7koiti	183	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	183	10	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	183	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	183	12	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	183	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	183	14	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	183	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	183	16	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	183	17	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	183	18	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	183	19	German	german	JJ
work_fklnz2kdpjhwfjysurpx7koiti	183	20	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	183	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	184	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	184	2	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	184	3	extends	extend	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	184	4	easily	easily	RB
work_fklnz2kdpjhwfjysurpx7koiti	184	5	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	184	6	various	various	JJ
work_fklnz2kdpjhwfjysurpx7koiti	184	7	foreign	foreign	JJ
work_fklnz2kdpjhwfjysurpx7koiti	184	8	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	184	9	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	184	10	de-	de-	RB
work_fklnz2kdpjhwfjysurpx7koiti	184	11	scribed	scribe	VBD
work_fklnz2kdpjhwfjysurpx7koiti	184	12	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	184	13	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	184	14	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	184	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	185	1	Another	another	DT
work_fklnz2kdpjhwfjysurpx7koiti	185	2	related	related	JJ
work_fklnz2kdpjhwfjysurpx7koiti	185	3	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	185	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	185	6	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	185	8	lexico-	lexico-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	9	syntactic	syntactic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	185	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	185	11	lexico	lexico	JJ
work_fklnz2kdpjhwfjysurpx7koiti	185	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	185	13	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	185	14	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	15	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	185	16	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	17	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	18	terns	tern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	185	19	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	185	20	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	185	21	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	22	corpora	corpora	NN
work_fklnz2kdpjhwfjysurpx7koiti	185	23	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	185	24	Hearst	Hearst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	25	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	185	26	1992	1992	CD
work_fklnz2kdpjhwfjysurpx7koiti	185	27	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	185	28	Chklovski	Chklovski	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	29	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	185	30	Pantel	Pantel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	185	32	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	185	33	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	185	34	Tandon	Tandon	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	35	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	185	36	de	de	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	37	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	185	38	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	185	39	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	185	40	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	185	41	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	186	1	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	186	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	186	3	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	186	4	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	186	5	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	186	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	186	7	follows	follow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	186	8	Davidov	Davidov	NNP
work_fklnz2kdpjhwfjysurpx7koiti	186	9	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	186	10	Rappoport	Rappoport	NNP
work_fklnz2kdpjhwfjysurpx7koiti	186	11	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	186	12	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	186	13	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	186	14	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	186	15	automatically	automatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	186	16	bootstrap	bootstrap	VB
work_fklnz2kdpjhwfjysurpx7koiti	186	17	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	186	18	scaling	scale	VBG
work_fklnz2kdpjhwfjysurpx7koiti	186	19	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	186	20	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	186	21	seed	seed	NN
work_fklnz2kdpjhwfjysurpx7koiti	186	22	ad-	ad-	CC
work_fklnz2kdpjhwfjysurpx7koiti	186	23	jectives	jective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	186	24	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	186	25	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	186	26	hits	hit	NNS
work_fklnz2kdpjhwfjysurpx7koiti	186	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	187	1	These	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	187	2	methods	method	NNS
work_fklnz2kdpjhwfjysurpx7koiti	187	3	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	187	4	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	187	5	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	187	6	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	187	7	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	187	8	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	187	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	187	10	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	187	11	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	187	12	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	187	13	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	187	14	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	187	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	188	1	VerbOcean	verbocean	JJ
work_fklnz2kdpjhwfjysurpx7koiti	188	2	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	188	3	Chklovski	Chklovski	NNP
work_fklnz2kdpjhwfjysurpx7koiti	188	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	188	5	Pantel	Pantel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	188	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	188	7	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	188	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	188	9	ex-	ex-	RB
work_fklnz2kdpjhwfjysurpx7koiti	188	10	tracts	tract	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	188	11	various	various	JJ
work_fklnz2kdpjhwfjysurpx7koiti	188	12	fine	fine	RB
work_fklnz2kdpjhwfjysurpx7koiti	188	13	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	188	14	grained	grain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	188	15	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	188	16	relations	relation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	188	17	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	188	18	in-	in-	XX
work_fklnz2kdpjhwfjysurpx7koiti	188	19	cluding	clude	VBG
work_fklnz2kdpjhwfjysurpx7koiti	188	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	188	21	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	188	22	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	188	23	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	188	24	relation	relation	NN
work_fklnz2kdpjhwfjysurpx7koiti	188	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	188	26	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	188	27	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	188	28	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	188	29	verbs	verb	NNS
work_fklnz2kdpjhwfjysurpx7koiti	188	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	188	31	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	188	32	lexico	lexico	NN
work_fklnz2kdpjhwfjysurpx7koiti	188	33	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	188	34	syntactic	syntactic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	188	35	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	188	36	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	188	37	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	188	38	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	188	39	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	189	1	283	283	CD
work_fklnz2kdpjhwfjysurpx7koiti	189	2	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	189	3	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	189	5	jointly	jointly	RB
work_fklnz2kdpjhwfjysurpx7koiti	189	6	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	189	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	189	8	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	189	10	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	189	11	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	189	12	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	13	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	14	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	189	15	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	189	16	applicable	applicable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	189	17	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	189	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	189	19	VerbO-	VerbO-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	189	20	cean	cean	NNP
work_fklnz2kdpjhwfjysurpx7koiti	189	21	pairs	pairs	NNP
work_fklnz2kdpjhwfjysurpx7koiti	189	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	189	23	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	189	24	should	should	MD
work_fklnz2kdpjhwfjysurpx7koiti	189	25	help	help	VB
work_fklnz2kdpjhwfjysurpx7koiti	189	26	address	address	VB
work_fklnz2kdpjhwfjysurpx7koiti	189	27	similar	similar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	189	28	sparsity	sparsity	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	29	issues	issue	NNS
work_fklnz2kdpjhwfjysurpx7koiti	189	30	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	189	31	local	local	JJ
work_fklnz2kdpjhwfjysurpx7koiti	189	32	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	189	33	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	189	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	190	1	Such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	190	2	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	190	3	will	will	MD
work_fklnz2kdpjhwfjysurpx7koiti	190	4	again	again	RB
work_fklnz2kdpjhwfjysurpx7koiti	190	5	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	190	6	quite	quite	RB
work_fklnz2kdpjhwfjysurpx7koiti	190	7	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	190	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	190	9	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	190	10	learners	learner	NNS
work_fklnz2kdpjhwfjysurpx7koiti	190	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	190	12	lan-	lan-	NN
work_fklnz2kdpjhwfjysurpx7koiti	190	13	guage	guage	NN
work_fklnz2kdpjhwfjysurpx7koiti	190	14	understanding	understanding	NN
work_fklnz2kdpjhwfjysurpx7koiti	190	15	tools	tool	NNS
work_fklnz2kdpjhwfjysurpx7koiti	190	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	191	1	de	de	NNP
work_fklnz2kdpjhwfjysurpx7koiti	191	2	Marneffe	Marneffe	NNP
work_fklnz2kdpjhwfjysurpx7koiti	191	3	et	et	NNP
work_fklnz2kdpjhwfjysurpx7koiti	191	4	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	191	5	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	192	1	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	192	2	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	192	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	192	4	infer	infer	VB
work_fklnz2kdpjhwfjysurpx7koiti	192	5	yes	yes	UH
work_fklnz2kdpjhwfjysurpx7koiti	192	6	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	192	7	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	192	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	192	9	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	192	10	answers	answer	NNS
work_fklnz2kdpjhwfjysurpx7koiti	192	11	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	192	12	questions	question	NNS
work_fklnz2kdpjhwfjysurpx7koiti	192	13	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	192	14	responses	response	NNS
work_fklnz2kdpjhwfjysurpx7koiti	192	15	involving	involve	VBG
work_fklnz2kdpjhwfjysurpx7koiti	192	16	scalar	scalar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	192	17	adjec-	adjec-	NN
work_fklnz2kdpjhwfjysurpx7koiti	192	18	tives	tive	NNS
work_fklnz2kdpjhwfjysurpx7koiti	192	19	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	192	20	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	192	21	dialogue	dialogue	NN
work_fklnz2kdpjhwfjysurpx7koiti	192	22	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	192	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	193	1	They	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	193	2	correlate	correlate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	193	3	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	193	4	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	193	5	ratings	rating	NNS
work_fklnz2kdpjhwfjysurpx7koiti	193	6	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	193	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	193	8	movie	movie	NN
work_fklnz2kdpjhwfjysurpx7koiti	193	9	review	review	NN
work_fklnz2kdpjhwfjysurpx7koiti	193	10	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	193	11	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	193	12	find	find	VB
work_fklnz2kdpjhwfjysurpx7koiti	193	13	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	193	14	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	193	15	appears	appear	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	193	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	193	17	lower	lower	RBR
work_fklnz2kdpjhwfjysurpx7koiti	193	18	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	193	19	rated	rate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	193	20	reviews	review	NNS
work_fklnz2kdpjhwfjysurpx7koiti	193	21	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	193	22	excellent	excellent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	193	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	194	1	Finally	finally	RB
work_fklnz2kdpjhwfjysurpx7koiti	194	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	3	there	there	EX
work_fklnz2kdpjhwfjysurpx7koiti	194	4	has	have	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	194	5	been	be	VBN
work_fklnz2kdpjhwfjysurpx7koiti	194	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	194	7	lot	lot	NN
work_fklnz2kdpjhwfjysurpx7koiti	194	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	194	9	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	194	10	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	194	11	measuring	measure	VBG
work_fklnz2kdpjhwfjysurpx7koiti	194	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	194	13	general	general	JJ
work_fklnz2kdpjhwfjysurpx7koiti	194	14	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	194	15	polarity	polarity	NN
work_fklnz2kdpjhwfjysurpx7koiti	194	16	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	194	17	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	194	18	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	194	19	Hatzivas-	Hatzivas-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	20	siloglou	siloglou	NN
work_fklnz2kdpjhwfjysurpx7koiti	194	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	22	McKeown	McKeown	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	24	1997	1997	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	25	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	26	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	27	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	28	Wiebe	Wiebe	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	30	2000	2000	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	31	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	32	Turney	Turney	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	33	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	34	Littman	Littman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	35	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	36	2003	2003	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	37	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	38	Liu	Liu	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	39	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	40	Seneff	Seneff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	41	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	42	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	43	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	44	Taboada	Taboada	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	45	et	et	FW
work_fklnz2kdpjhwfjysurpx7koiti	194	46	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	47	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	48	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	49	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	50	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	51	Yessenalina	Yessenalina	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	52	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	53	Cardie	Cardie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	54	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	55	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	56	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	194	57	Pang	Pang	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	58	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	194	59	Lee	Lee	NNP
work_fklnz2kdpjhwfjysurpx7koiti	194	60	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	194	61	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	194	62	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	194	63	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	195	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	195	2	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	3	in-	in-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	195	4	stead	stead	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	5	aims	aim	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	195	6	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	7	producing	produce	VBG
work_fklnz2kdpjhwfjysurpx7koiti	195	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	195	9	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	195	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	195	11	unrestricted	unrestricted	JJ
work_fklnz2kdpjhwfjysurpx7koiti	195	12	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	14	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	195	15	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	16	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	195	17	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	18	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	195	19	qualities	quality	NNS
work_fklnz2kdpjhwfjysurpx7koiti	195	20	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	195	21	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	195	22	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	195	23	help	help	VB
work_fklnz2kdpjhwfjysurpx7koiti	195	24	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	25	fine	fine	RB
work_fklnz2kdpjhwfjysurpx7koiti	195	26	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	195	27	grained	grain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	195	28	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	29	analy-	analy-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	195	30	sis	sis	NNP
work_fklnz2kdpjhwfjysurpx7koiti	195	31	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	32	respect	respect	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	33	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	195	34	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	195	35	particular	particular	JJ
work_fklnz2kdpjhwfjysurpx7koiti	195	36	content	content	NN
work_fklnz2kdpjhwfjysurpx7koiti	195	37	aspects	aspect	NNS
work_fklnz2kdpjhwfjysurpx7koiti	195	38	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	196	1	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	196	2	Experiments	experiment	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	3	4.1	4.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	196	4	Data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	5	Input	Input	NNP
work_fklnz2kdpjhwfjysurpx7koiti	196	6	Clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	7	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	8	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	9	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	196	10	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	196	11	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	12	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	13	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	14	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	196	16	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	196	17	started	start	VBD
work_fklnz2kdpjhwfjysurpx7koiti	196	18	out	out	RP
work_fklnz2kdpjhwfjysurpx7koiti	196	19	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	196	21	satellite	satellite	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	22	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	23	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	196	24	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	196	25	dumbbell	dumbbell	NNP
work_fklnz2kdpjhwfjysurpx7koiti	196	26	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	196	27	structure	structure	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	28	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	29	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	30	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	31	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	196	32	3.0	3.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	196	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	196	34	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	196	35	consists	consist	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	196	36	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	37	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	196	38	direct	direct	JJ
work_fklnz2kdpjhwfjysurpx7koiti	196	39	antonyms	antonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	40	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	41	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	196	42	poles	pole	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	43	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	196	44	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	196	45	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	46	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	47	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	196	48	satellite	satellite	NN
work_fklnz2kdpjhwfjysurpx7koiti	196	49	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	50	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	196	51	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	196	52	se-	se-	VBN
work_fklnz2kdpjhwfjysurpx7koiti	196	53	mantically	mantically	RB
work_fklnz2kdpjhwfjysurpx7koiti	196	54	similar	similar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	196	55	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	56	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	196	57	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	196	58	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	196	59	poles	pole	NNS
work_fklnz2kdpjhwfjysurpx7koiti	196	60	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	196	61	Gross	Gross	NNP
work_fklnz2kdpjhwfjysurpx7koiti	196	62	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	196	63	Miller	Miller	NNP
work_fklnz2kdpjhwfjysurpx7koiti	196	64	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	196	65	1990	1990	CD
work_fklnz2kdpjhwfjysurpx7koiti	196	66	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	196	67	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	197	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	197	2	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	197	3	antonymy	antonymy	JJ
work_fklnz2kdpjhwfjysurpx7koiti	197	4	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	197	6	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	197	7	deter-	deter-	VBD
work_fklnz2kdpjhwfjysurpx7koiti	197	8	mined	mine	VBD
work_fklnz2kdpjhwfjysurpx7koiti	197	9	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	197	10	extended	extended	JJ
work_fklnz2kdpjhwfjysurpx7koiti	197	11	dumbbell	dumbbell	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	12	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	197	13	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	197	14	looking	look	VBG
work_fklnz2kdpjhwfjysurpx7koiti	197	15	up	up	RP
work_fklnz2kdpjhwfjysurpx7koiti	197	16	syn-	syn-	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	17	onyms	onyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	18	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	197	19	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	197	20	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	197	21	related	related	JJ
work_fklnz2kdpjhwfjysurpx7koiti	197	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	197	23	satellite	satellite	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	24	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	25	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	197	26	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	197	27	see	see	VB
work_fklnz2kdpjhwfjysurpx7koiti	197	28	-	-	:
work_fklnz2kdpjhwfjysurpx7koiti	197	29	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	197	30	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	197	31	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	197	32	synonym	synonym	NN
work_fklnz2kdpjhwfjysurpx7koiti	197	33	sets	set	NNS
work_fklnz2kdpjhwfjysurpx7koiti	197	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	198	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	198	2	cut	cut	VBD
work_fklnz2kdpjhwfjysurpx7koiti	198	3	such	such	PDT
work_fklnz2kdpjhwfjysurpx7koiti	198	4	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	198	5	extended	extended	JJ
work_fklnz2kdpjhwfjysurpx7koiti	198	6	dumbbell	dumbbell	NN
work_fklnz2kdpjhwfjysurpx7koiti	198	7	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	198	8	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	198	9	antonymous	antonymous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	198	10	halves	half	NNS
work_fklnz2kdpjhwfjysurpx7koiti	198	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	198	12	treated	treat	VBD
work_fklnz2kdpjhwfjysurpx7koiti	198	13	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	198	14	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	198	15	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	198	16	halves	half	NNS
work_fklnz2kdpjhwfjysurpx7koiti	198	17	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	198	18	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	198	19	potential	potential	JJ
work_fklnz2kdpjhwfjysurpx7koiti	198	20	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	198	21	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	198	22	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	198	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	199	1	Most	Most	JJS
work_fklnz2kdpjhwfjysurpx7koiti	199	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	199	3	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	199	4	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	199	5	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	199	6	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	199	7	noisy	noisy	JJ
work_fklnz2kdpjhwfjysurpx7koiti	199	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	199	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	199	10	purpose	purpose	NN
work_fklnz2kdpjhwfjysurpx7koiti	199	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	199	12	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	199	13	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	199	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	199	15	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	200	1	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	200	2	contain	contain	VBP
work_fklnz2kdpjhwfjysurpx7koiti	200	3	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	200	4	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	200	5	appear	appear	VBP
work_fklnz2kdpjhwfjysurpx7koiti	200	6	unrelatable	unrelatable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	200	7	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	200	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	200	9	single	single	JJ
work_fklnz2kdpjhwfjysurpx7koiti	200	10	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	200	11	due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	200	12	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	200	13	polysemy	polysemy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	200	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	200	15	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	200	16	drift	drift	NN
work_fklnz2kdpjhwfjysurpx7koiti	200	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	200	18	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	201	1	violent	violent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	201	2	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	201	3	respect	respect	NN
work_fklnz2kdpjhwfjysurpx7koiti	201	4	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	201	5	super-	super-	NN
work_fklnz2kdpjhwfjysurpx7koiti	201	6	natural	natural	JJ
work_fklnz2kdpjhwfjysurpx7koiti	201	7	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	201	8	affected	affected	JJ
work_fklnz2kdpjhwfjysurpx7koiti	201	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	202	1	Motivated	motivate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	202	2	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	202	3	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	202	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	202	5	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	202	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	202	7	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	202	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	202	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	202	10	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	202	11	split	split	VBD
work_fklnz2kdpjhwfjysurpx7koiti	202	12	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	202	13	hard	hard	RB
work_fklnz2kdpjhwfjysurpx7koiti	202	14	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	202	15	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	202	16	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	202	17	relate	relate	NN
work_fklnz2kdpjhwfjysurpx7koiti	202	18	ad-	ad-	CC
work_fklnz2kdpjhwfjysurpx7koiti	202	19	jectives	jective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	202	20	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	202	21	smaller	small	JJR
work_fklnz2kdpjhwfjysurpx7koiti	202	22	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	202	23	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	202	24	specific	specific	JJ
work_fklnz2kdpjhwfjysurpx7koiti	202	25	subgroups	subgroup	NNS
work_fklnz2kdpjhwfjysurpx7koiti	202	26	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	202	27	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	202	28	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	202	29	evidence4	evidence4	NN
work_fklnz2kdpjhwfjysurpx7koiti	202	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	203	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	203	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	203	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	203	5	consider	consider	VBP
work_fklnz2kdpjhwfjysurpx7koiti	203	6	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	203	7	undi-	undi-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	203	8	4Note	4note	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	9	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	10	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	203	11	do	do	VBP
work_fklnz2kdpjhwfjysurpx7koiti	203	12	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	203	13	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	203	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	203	15	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	203	16	dataset	dataset	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	17	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	18	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	203	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	203	20	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	203	21	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	203	22	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	203	24	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	25	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	203	27	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	28	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	203	29	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	203	30	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	203	31	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	203	32	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	203	33	438	438	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	34	115	115	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	35	60	60	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	36	35	35	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	37	19	19	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	38	12	12	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	39	14	14	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	40	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	41	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	42	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	43	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	44	100	100	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	45	200	200	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	46	300	300	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	47	400	400	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	48	500	500	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	49	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	50	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	51	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	52	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	53	6	6	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	54	7	7	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	55	8	8	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	56	9	9	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	57	10	10	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	58	-	-	SYM
work_fklnz2kdpjhwfjysurpx7koiti	203	59	14	14	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	60	15	15	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	61	-	-	SYM
work_fklnz2kdpjhwfjysurpx7koiti	203	62	17	17	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	63	#	#	$
work_fklnz2kdpjhwfjysurpx7koiti	203	64	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	65	c	c	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	66	ha	ha	UH
work_fklnz2kdpjhwfjysurpx7koiti	203	67	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	68	s	s	NNP
work_fklnz2kdpjhwfjysurpx7koiti	203	69	Length	Length	NNP
work_fklnz2kdpjhwfjysurpx7koiti	203	70	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	71	chain	chain	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	72	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	73	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	203	74	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	203	75	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	203	76	histogram	histogram	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	77	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	78	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	203	79	sizes	size	NNS
work_fklnz2kdpjhwfjysurpx7koiti	203	80	after	after	IN
work_fklnz2kdpjhwfjysurpx7koiti	203	81	partitioning	partition	VBG
work_fklnz2kdpjhwfjysurpx7koiti	203	82	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	204	1	41	41	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	2	27	27	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	3	12	12	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	4	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	5	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	6	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	7	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	8	10	10	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	9	20	20	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	10	30	30	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	11	40	40	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	12	50	50	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	13	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	14	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	15	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	16	6	6	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	17	7	7	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	18	8	8	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	19	#	#	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	204	21	c	c	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	22	ha	ha	UH
work_fklnz2kdpjhwfjysurpx7koiti	204	23	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	204	24	s	s	NNP
work_fklnz2kdpjhwfjysurpx7koiti	204	25	Length	Length	NNP
work_fklnz2kdpjhwfjysurpx7koiti	204	26	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	204	27	chain	chain	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	28	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	29	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	204	30	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	204	31	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	204	32	histogram	histogram	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	33	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	204	34	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	35	sizes	size	NNS
work_fklnz2kdpjhwfjysurpx7koiti	204	36	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	204	37	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	204	38	test	test	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	39	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	204	40	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	205	1	rected	recte	VBN
work_fklnz2kdpjhwfjysurpx7koiti	205	2	edge	edge	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	3	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	205	4	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	205	5	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	205	7	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	205	8	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	205	9	has	have	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	205	10	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	205	11	non	non	JJ
work_fklnz2kdpjhwfjysurpx7koiti	205	12	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	205	13	zero	zero	CD
work_fklnz2kdpjhwfjysurpx7koiti	205	14	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	15	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	205	17	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	205	18	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	205	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	205	20	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	21	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	205	22	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	23	scoring	score	VBG
work_fklnz2kdpjhwfjysurpx7koiti	205	24	procedure	procedure	NN
work_fklnz2kdpjhwfjysurpx7koiti	205	25	described	describe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	205	26	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	205	27	Section	Section	NNP
work_fklnz2kdpjhwfjysurpx7koiti	205	28	2.1.3	2.1.3	CD
work_fklnz2kdpjhwfjysurpx7koiti	205	29	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	205	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	206	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	206	2	resulting	result	VBG
work_fklnz2kdpjhwfjysurpx7koiti	206	3	graph	graph	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	206	5	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	206	6	partitioned	partition	VBN
work_fklnz2kdpjhwfjysurpx7koiti	206	7	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	8	connected	connected	JJ
work_fklnz2kdpjhwfjysurpx7koiti	206	9	components	component	NNS
work_fklnz2kdpjhwfjysurpx7koiti	206	10	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	206	11	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	12	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	206	13	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	206	14	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	15	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	206	16	subgraph	subgraph	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	17	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	206	18	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	19	least	least	JJS
work_fklnz2kdpjhwfjysurpx7koiti	206	20	indirectly	indirectly	RB
work_fklnz2kdpjhwfjysurpx7koiti	206	21	connected	connect	VBN
work_fklnz2kdpjhwfjysurpx7koiti	206	22	via	via	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	23	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	206	24	path	path	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	25	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	206	26	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	206	27	much	much	RB
work_fklnz2kdpjhwfjysurpx7koiti	206	28	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	206	29	likely	likely	JJ
work_fklnz2kdpjhwfjysurpx7koiti	206	30	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	206	31	belong	belong	VB
work_fklnz2kdpjhwfjysurpx7koiti	206	32	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	206	33	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	206	34	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	206	35	inten-	inten-	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	36	sity	sity	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	37	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	206	38	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	207	1	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	207	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	207	3	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	207	4	break	break	VB
work_fklnz2kdpjhwfjysurpx7koiti	207	5	up	up	RP
work_fklnz2kdpjhwfjysurpx7koiti	207	6	partitions	partition	NNS
work_fklnz2kdpjhwfjysurpx7koiti	207	7	when-	when-	VB
work_fklnz2kdpjhwfjysurpx7koiti	207	8	ever	ever	RB
work_fklnz2kdpjhwfjysurpx7koiti	207	9	there	there	EX
work_fklnz2kdpjhwfjysurpx7koiti	207	10	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	207	11	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	207	12	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	207	13	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	207	14	connecting	connect	VBG
work_fklnz2kdpjhwfjysurpx7koiti	207	15	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	207	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	207	17	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	207	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	207	19	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	207	20	within	within	IN
work_fklnz2kdpjhwfjysurpx7koiti	207	21	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	207	22	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	207	23	partition	partition	NN
work_fklnz2kdpjhwfjysurpx7koiti	207	24	re-	re-	RB
work_fklnz2kdpjhwfjysurpx7koiti	207	25	mains	main	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	207	26	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	207	27	challenging	challenging	JJ
work_fklnz2kdpjhwfjysurpx7koiti	207	28	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	207	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	208	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	208	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	208	3	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	208	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	208	5	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	208	6	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	208	7	will	will	MD
work_fklnz2kdpjhwfjysurpx7koiti	208	8	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	208	9	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	208	10	necessarily	necessarily	RB
work_fklnz2kdpjhwfjysurpx7koiti	208	11	directly	directly	RB
work_fklnz2kdpjhwfjysurpx7koiti	208	12	relate	relate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	208	13	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	208	14	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	208	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	208	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	208	17	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	208	18	partition	partition	NN
work_fklnz2kdpjhwfjysurpx7koiti	208	19	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	208	20	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	208	21	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	208	22	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	208	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	209	1	Addition-	Addition-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	209	2	ally	ally	NN
work_fklnz2kdpjhwfjysurpx7koiti	209	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	209	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	209	5	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	209	6	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	209	7	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	209	8	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	209	9	indicate	indicate	VB
work_fklnz2kdpjhwfjysurpx7koiti	209	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	209	11	wrong	wrong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	209	12	direction	direction	NN
work_fklnz2kdpjhwfjysurpx7koiti	209	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	210	1	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	210	2	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	210	3	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	210	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	210	5	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	210	6	distribution	distribution	NN
work_fklnz2kdpjhwfjysurpx7koiti	210	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	210	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	210	9	resulting	result	VBG
work_fklnz2kdpjhwfjysurpx7koiti	210	10	partitions	partition	NNS
work_fklnz2kdpjhwfjysurpx7koiti	210	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	211	1	Patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	2	To	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	211	3	construct	construct	VB
work_fklnz2kdpjhwfjysurpx7koiti	211	4	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	211	5	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	6	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	7	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	211	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	211	9	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	211	10	started	start	VBD
work_fklnz2kdpjhwfjysurpx7koiti	211	11	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	12	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	211	13	couple	couple	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	14	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	15	common	common	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	16	rankable	rankable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	17	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	18	seed	seed	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	19	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	20	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	21	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	23	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	211	25	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	27	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	211	28	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	29	hot	hot	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	211	31	boiling	boil	VBG
work_fklnz2kdpjhwfjysurpx7koiti	211	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	33	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	211	34	used	use	VBD
work_fklnz2kdpjhwfjysurpx7koiti	211	35	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	211	36	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	37	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	211	38	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	39	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	211	40	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	211	41	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	42	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	211	43	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	44	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	211	45	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	211	46	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	211	47	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	211	48	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	211	49	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	50	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	211	51	collect	collect	VB
work_fklnz2kdpjhwfjysurpx7koiti	211	52	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	211	53	few	few	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	54	most	most	RBS
work_fklnz2kdpjhwfjysurpx7koiti	211	55	frequent	frequent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	211	56	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	57	terns	tern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	58	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	59	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	211	60	around	around	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	61	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	211	62	seed	seed	NN
work_fklnz2kdpjhwfjysurpx7koiti	211	63	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	211	64	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	65	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	66	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	211	67	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	211	68	directions	direction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	211	69	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	211	70	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	212	1	Among	among	IN
work_fklnz2kdpjhwfjysurpx7koiti	212	2	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	212	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	212	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	212	5	manually	manually	RB
work_fklnz2kdpjhwfjysurpx7koiti	212	6	chose	choose	VBD
work_fklnz2kdpjhwfjysurpx7koiti	212	7	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	212	8	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	212	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	213	1	Instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	213	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	3	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	213	4	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	213	5	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	213	6	made	make	VBD
work_fklnz2kdpjhwfjysurpx7koiti	213	7	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	213	8	compar-	compar-	NN
work_fklnz2kdpjhwfjysurpx7koiti	213	9	isons	ison	NNS
work_fklnz2kdpjhwfjysurpx7koiti	213	10	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	213	11	select	select	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	12	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	213	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	14	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	213	15	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	213	16	5-way	5-way	NNP
work_fklnz2kdpjhwfjysurpx7koiti	213	17	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	213	18	scheme	scheme	NN
work_fklnz2kdpjhwfjysurpx7koiti	213	19	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	213	20	neutral	neutral	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	22	mild	mild	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	24	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	213	25	mild	mild	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	27	intense	intense	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	213	29	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	213	30	intense	intense	JJ
work_fklnz2kdpjhwfjysurpx7koiti	213	31	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	213	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	214	1	284	284	CD
work_fklnz2kdpjhwfjysurpx7koiti	214	2	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	214	3	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	214	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	214	5	intuitive	intuitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	214	6	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	214	7	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	214	8	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	214	9	linguistically	linguistically	RB
work_fklnz2kdpjhwfjysurpx7koiti	214	10	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	214	11	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	214	12	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	214	13	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	214	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	214	15	several	several	JJ
work_fklnz2kdpjhwfjysurpx7koiti	214	16	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	214	17	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	214	18	had	have	VBD
work_fklnz2kdpjhwfjysurpx7koiti	214	19	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	214	20	been	be	VBN
work_fklnz2kdpjhwfjysurpx7koiti	214	21	discovered	discover	VBN
work_fklnz2kdpjhwfjysurpx7koiti	214	22	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	214	23	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	214	24	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	214	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	215	1	These	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	215	2	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	215	3	shown	show	VBN
work_fklnz2kdpjhwfjysurpx7koiti	215	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	215	5	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	215	6	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	215	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	216	1	Note	note	VB
work_fklnz2kdpjhwfjysurpx7koiti	216	2	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	216	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	216	4	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	216	5	collected	collect	VBD
work_fklnz2kdpjhwfjysurpx7koiti	216	6	pat-	pat-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	216	7	terns	tern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	216	8	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	216	9	were	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	216	10	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	216	11	ambiguous	ambiguous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	216	12	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	216	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	216	14	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	216	15	orders	order	NNS
work_fklnz2kdpjhwfjysurpx7koiti	216	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	216	17	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	216	18	example	example	NN
work_fklnz2kdpjhwfjysurpx7koiti	216	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	216	20	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	216	21	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	216	22	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	217	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	217	2	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	217	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	217	4	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	217	5	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	217	6	ambiguous	ambiguous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	217	7	be-	be-	XX
work_fklnz2kdpjhwfjysurpx7koiti	217	8	cause	cause	IN
work_fklnz2kdpjhwfjysurpx7koiti	217	9	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	217	10	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	217	11	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	217	12	used	use	VBN
work_fklnz2kdpjhwfjysurpx7koiti	217	13	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	217	14	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	217	15	’	'	``
work_fklnz2kdpjhwfjysurpx7koiti	217	16	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	217	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	217	18	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	217	19	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	217	20	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	217	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	217	22	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	217	23	great	great	JJ
work_fklnz2kdpjhwfjysurpx7koiti	217	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	217	25	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	217	26	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	217	27	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	217	28	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	218	1	Alternatively	alternatively	RB
work_fklnz2kdpjhwfjysurpx7koiti	218	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	3	one	one	PRP
work_fklnz2kdpjhwfjysurpx7koiti	218	4	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	218	5	easily	easily	RB
work_fklnz2kdpjhwfjysurpx7koiti	218	6	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	218	7	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	218	8	fully	fully	RB
work_fklnz2kdpjhwfjysurpx7koiti	218	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	218	10	automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	218	11	bootstrapping	bootstrapping	NN
work_fklnz2kdpjhwfjysurpx7koiti	218	12	techniques	technique	NNS
work_fklnz2kdpjhwfjysurpx7koiti	218	13	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	218	14	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	218	15	seed	seed	NN
work_fklnz2kdpjhwfjysurpx7koiti	218	16	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	218	17	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	218	18	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	218	19	Hearst	Hearst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	21	1992	1992	CD
work_fklnz2kdpjhwfjysurpx7koiti	218	22	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	218	23	Chklovski	Chklovski	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	24	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	218	25	Pantel	Pantel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	26	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	27	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	218	28	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	218	29	Yang	Yang	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	30	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	218	31	Su	Su	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	32	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	33	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	218	34	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	218	35	Turney	Turney	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	36	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	37	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	218	38	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	218	39	Davidov	Davidov	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	40	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	218	41	Rappoport	Rappoport	NNP
work_fklnz2kdpjhwfjysurpx7koiti	218	42	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	218	43	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	218	44	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	218	45	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	219	1	However	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	219	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	219	3	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	219	4	semi-	semi-	NN
work_fklnz2kdpjhwfjysurpx7koiti	219	5	automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	6	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	219	7	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	219	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	219	9	simple	simple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	219	11	fast	fast	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	12	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	219	13	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	219	14	extracts	extract	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	219	15	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	219	16	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	17	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	219	18	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	219	19	high	high	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	20	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	219	21	quality	quality	NN
work_fklnz2kdpjhwfjysurpx7koiti	219	22	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	219	23	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	219	24	gen-	gen-	RB
work_fklnz2kdpjhwfjysurpx7koiti	219	25	eral	eral	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	26	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	219	27	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	219	28	scaling	scale	VBG
work_fklnz2kdpjhwfjysurpx7koiti	219	29	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	219	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	220	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	220	2	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	220	3	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	220	4	quickly	quickly	RB
work_fklnz2kdpjhwfjysurpx7koiti	220	5	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	220	6	repeated	repeat	VBN
work_fklnz2kdpjhwfjysurpx7koiti	220	7	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	220	8	scratch	scratch	NN
work_fklnz2kdpjhwfjysurpx7koiti	220	9	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	220	10	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	220	11	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	220	12	lan-	lan-	NN
work_fklnz2kdpjhwfjysurpx7koiti	220	13	guage	guage	NN
work_fklnz2kdpjhwfjysurpx7koiti	220	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	221	1	Moreover	moreover	RB
work_fklnz2kdpjhwfjysurpx7koiti	221	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	221	3	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	221	4	described	describe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	221	5	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	221	6	Section	Section	NNP
work_fklnz2kdpjhwfjysurpx7koiti	221	7	5.1	5.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	221	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	221	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	221	10	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	221	11	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	221	12	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	221	13	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	221	14	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	221	15	projected	project	VBN
work_fklnz2kdpjhwfjysurpx7koiti	221	16	automatically	automatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	221	17	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	221	18	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	221	19	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	221	20	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	221	21	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	221	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	222	1	Development	development	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	222	3	Test	Test	NNP
work_fklnz2kdpjhwfjysurpx7koiti	222	4	Sets	Sets	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	222	5	Section	Section	NNP
work_fklnz2kdpjhwfjysurpx7koiti	222	6	2.1	2.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	222	7	describes	describe	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	222	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	222	9	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	10	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	222	11	collecting	collect	VBG
work_fklnz2kdpjhwfjysurpx7koiti	222	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	222	13	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	14	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	222	15	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	222	16	ad-	ad-	CC
work_fklnz2kdpjhwfjysurpx7koiti	222	17	jective	jective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	222	18	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	222	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	222	20	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	222	21	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	22	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	222	23	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	24	n	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	222	25	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	222	26	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	222	27	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	222	28	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	222	29	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	222	30	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	222	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	222	32	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	222	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	222	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	223	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	223	2	relied	rely	VBD
work_fklnz2kdpjhwfjysurpx7koiti	223	3	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	223	4	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	223	5	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	223	6	development	development	NN
work_fklnz2kdpjhwfjysurpx7koiti	223	7	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	223	8	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	223	9	test	test	VB
work_fklnz2kdpjhwfjysurpx7koiti	223	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	223	11	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	223	12	structure	structure	NN
work_fklnz2kdpjhwfjysurpx7koiti	223	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	223	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	223	15	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	223	16	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	223	17	setup	setup	NN
work_fklnz2kdpjhwfjysurpx7koiti	223	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	224	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	224	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	224	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	224	4	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	224	5	manually	manually	RB
work_fklnz2kdpjhwfjysurpx7koiti	224	6	chose	choose	VBD
work_fklnz2kdpjhwfjysurpx7koiti	224	7	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	224	8	representative	representative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	224	9	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	224	10	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	224	11	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	224	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	224	13	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	224	14	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	224	15	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	224	16	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	224	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	225	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	225	2	final	final	JJ
work_fklnz2kdpjhwfjysurpx7koiti	225	3	test	test	NN
work_fklnz2kdpjhwfjysurpx7koiti	225	4	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	225	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	225	6	distinct	distinct	JJ
work_fklnz2kdpjhwfjysurpx7koiti	225	7	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	225	8	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	225	9	development	development	NN
work_fklnz2kdpjhwfjysurpx7koiti	225	10	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	225	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	225	12	consists	consist	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	225	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	225	14	569	569	CD
work_fklnz2kdpjhwfjysurpx7koiti	225	15	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	225	16	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	225	17	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	225	18	88	88	CD
work_fklnz2kdpjhwfjysurpx7koiti	225	19	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	225	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	225	21	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	225	22	annotated	annotate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	225	23	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	225	24	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	225	25	native	native	JJ
work_fklnz2kdpjhwfjysurpx7koiti	225	26	speakers	speaker	NNS
work_fklnz2kdpjhwfjysurpx7koiti	225	27	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	225	28	English	English	NNP
work_fklnz2kdpjhwfjysurpx7koiti	225	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	226	1	Both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	226	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	226	3	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	226	4	test	test	NN
work_fklnz2kdpjhwfjysurpx7koiti	226	5	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	226	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	226	7	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	226	8	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	226	9	code	code	NN
work_fklnz2kdpjhwfjysurpx7koiti	226	10	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	226	11	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	226	12	freely	freely	RB
work_fklnz2kdpjhwfjysurpx7koiti	226	13	avail-	avail-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	226	14	able.5	able.5	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	15	To	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	226	16	arrive	arrive	VB
work_fklnz2kdpjhwfjysurpx7koiti	226	17	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	226	18	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	226	19	data	data	NN
work_fklnz2kdpjhwfjysurpx7koiti	226	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	226	21	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	226	22	randomly	randomly	RB
work_fklnz2kdpjhwfjysurpx7koiti	226	23	drew	draw	VBD
work_fklnz2kdpjhwfjysurpx7koiti	226	24	30	30	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	25	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	226	26	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	226	27	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	226	28	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	226	29	sizes	size	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	226	30	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	226	32	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	226	34	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	226	35	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	36	+	+	SYM
work_fklnz2kdpjhwfjysurpx7koiti	226	37	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	226	38	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	226	39	histogram	histogram	NN
work_fklnz2kdpjhwfjysurpx7koiti	226	40	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	226	41	partitioned	partition	VBN
work_fklnz2kdpjhwfjysurpx7koiti	226	42	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	226	43	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	226	44	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	226	45	Fig-	Fig-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	226	46	ure	ure	NNP
work_fklnz2kdpjhwfjysurpx7koiti	226	47	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	226	48	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	227	1	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	227	2	labeling	label	VBG
work_fklnz2kdpjhwfjysurpx7koiti	227	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	227	4	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	227	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	227	6	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	227	7	could	could	MD
work_fklnz2kdpjhwfjysurpx7koiti	227	8	ex-	ex-	RB
work_fklnz2kdpjhwfjysurpx7koiti	227	9	clude	clude	VB
work_fklnz2kdpjhwfjysurpx7koiti	227	10	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	227	11	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	227	12	they	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	227	13	deemed	deem	VBD
work_fklnz2kdpjhwfjysurpx7koiti	227	14	unsuitable	unsuitable	JJ
work_fklnz2kdpjhwfjysurpx7koiti	227	15	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	227	16	fit	fit	VB
work_fklnz2kdpjhwfjysurpx7koiti	227	17	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	227	18	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	227	19	single	single	JJ
work_fklnz2kdpjhwfjysurpx7koiti	227	20	shared	share	VBN
work_fklnz2kdpjhwfjysurpx7koiti	227	21	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	227	22	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	227	23	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	227	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	227	25	rest	rest	NN
work_fklnz2kdpjhwfjysurpx7koiti	227	26	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	227	27	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	227	28	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	227	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	228	1	Fortunately	fortunately	RB
work_fklnz2kdpjhwfjysurpx7koiti	228	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	228	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	228	4	partitioning	partitioning	NN
work_fklnz2kdpjhwfjysurpx7koiti	228	5	described	describe	VBD
work_fklnz2kdpjhwfjysurpx7koiti	228	6	ear-	ear-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	228	7	lier	lier	NNP
work_fklnz2kdpjhwfjysurpx7koiti	228	8	had	have	VBD
work_fklnz2kdpjhwfjysurpx7koiti	228	9	already	already	RB
work_fklnz2kdpjhwfjysurpx7koiti	228	10	separated	separate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	228	11	most	most	JJS
work_fklnz2kdpjhwfjysurpx7koiti	228	12	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	228	13	cases	case	NNS
work_fklnz2kdpjhwfjysurpx7koiti	228	14	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	228	15	dis-	dis-	IN
work_fklnz2kdpjhwfjysurpx7koiti	228	16	tinct	tinct	NN
work_fklnz2kdpjhwfjysurpx7koiti	228	17	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	228	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	229	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	229	2	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	229	3	ordered	order	VBD
work_fklnz2kdpjhwfjysurpx7koiti	229	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	229	5	remaining	remain	VBG
work_fklnz2kdpjhwfjysurpx7koiti	229	6	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	229	7	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	229	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	229	9	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	229	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	230	1	Words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	230	2	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	230	3	seemed	seem	VBD
work_fklnz2kdpjhwfjysurpx7koiti	230	4	indistinguish-	indistinguish-	NN
work_fklnz2kdpjhwfjysurpx7koiti	230	5	able	able	JJ
work_fklnz2kdpjhwfjysurpx7koiti	230	6	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	230	7	strength	strength	NN
work_fklnz2kdpjhwfjysurpx7koiti	230	8	could	could	MD
work_fklnz2kdpjhwfjysurpx7koiti	230	9	share	share	VB
work_fklnz2kdpjhwfjysurpx7koiti	230	10	positions	position	NNS
work_fklnz2kdpjhwfjysurpx7koiti	230	11	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	230	12	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	230	13	anno-	anno-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	230	14	tation	tation	NN
work_fklnz2kdpjhwfjysurpx7koiti	230	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	231	1	As	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	231	2	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	231	3	goal	goal	NN
work_fklnz2kdpjhwfjysurpx7koiti	231	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	231	5	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	231	6	compare	compare	VB
work_fklnz2kdpjhwfjysurpx7koiti	231	7	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	231	8	formation	formation	NN
work_fklnz2kdpjhwfjysurpx7koiti	231	9	algo-	algo-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	231	10	rithms	rithms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	231	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	231	12	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	231	13	did	do	VBD
work_fklnz2kdpjhwfjysurpx7koiti	231	14	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	231	15	include	include	VB
work_fklnz2kdpjhwfjysurpx7koiti	231	16	trivial	trivial	JJ
work_fklnz2kdpjhwfjysurpx7koiti	231	17	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	231	18	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	231	19	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	231	20	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	231	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	232	1	On	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	2	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	232	3	trivial	trivial	JJ
work_fklnz2kdpjhwfjysurpx7koiti	232	4	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	232	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	232	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	7	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	8	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	9	alone	alone	RB
work_fklnz2kdpjhwfjysurpx7koiti	232	10	de-	de-	RB
work_fklnz2kdpjhwfjysurpx7koiti	232	11	termines	termine	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	232	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	13	output	output	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	232	15	hence	hence	RB
work_fklnz2kdpjhwfjysurpx7koiti	232	16	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	17	algorithms	algorithm	NNS
work_fklnz2kdpjhwfjysurpx7koiti	232	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	232	19	includ-	includ-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	232	20	5http://demelo.org/gdm/intensity/	5http://demelo.org/gdm/intensity/	NNP
work_fklnz2kdpjhwfjysurpx7koiti	232	21	ing	e	VBG
work_fklnz2kdpjhwfjysurpx7koiti	232	22	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	23	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	232	25	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	232	26	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	27	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	232	28	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	29	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	30	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	232	31	defined	define	VBN
work_fklnz2kdpjhwfjysurpx7koiti	232	32	below	below	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	232	34	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	35	93.3	93.3	CD
work_fklnz2kdpjhwfjysurpx7koiti	232	36	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	37	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	38	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	232	39	separate	separate	JJ
work_fklnz2kdpjhwfjysurpx7koiti	232	40	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	41	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	42	30	30	CD
work_fklnz2kdpjhwfjysurpx7koiti	232	43	ran-	ran-	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	44	dom	dom	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	45	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	232	46	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	232	47	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	232	48	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	232	49	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	233	1	Figure	figure	NN
work_fklnz2kdpjhwfjysurpx7koiti	233	2	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	233	3	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	233	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	233	5	distribution	distribution	NN
work_fklnz2kdpjhwfjysurpx7koiti	233	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	233	7	cluster	cluster	NN
work_fklnz2kdpjhwfjysurpx7koiti	233	8	sizes	size	NNS
work_fklnz2kdpjhwfjysurpx7koiti	233	9	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	233	10	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	233	11	main	main	JJ
work_fklnz2kdpjhwfjysurpx7koiti	233	12	gold	gold	NN
work_fklnz2kdpjhwfjysurpx7koiti	233	13	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	233	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	234	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	234	2	inter	inter	JJ
work_fklnz2kdpjhwfjysurpx7koiti	234	3	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	234	4	annotator	annotator	JJ
work_fklnz2kdpjhwfjysurpx7koiti	234	5	agreement	agreement	NN
work_fklnz2kdpjhwfjysurpx7koiti	234	6	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	234	7	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	234	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	234	9	Cohen	Cohen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	234	10	’s	’s	NNP
work_fklnz2kdpjhwfjysurpx7koiti	234	11	κ	κ	NN
work_fklnz2kdpjhwfjysurpx7koiti	234	12	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	234	13	Cohen	Cohen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	234	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	234	15	1960	1960	CD
work_fklnz2kdpjhwfjysurpx7koiti	234	16	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	234	17	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	234	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	234	19	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	234	20	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	234	21	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	234	22	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	234	23	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	234	24	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	234	25	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	234	26	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	234	27	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	234	28	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	234	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	234	30	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	234	31	equal	equal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	234	32	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	234	33	unknown	unknown	JJ
work_fklnz2kdpjhwfjysurpx7koiti	234	34	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	234	35	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	234	36	0.64	0.64	CD
work_fklnz2kdpjhwfjysurpx7koiti	234	37	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	235	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	235	2	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	235	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	235	4	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	235	5	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	235	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	235	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	235	8	agreement	agreement	NN
work_fklnz2kdpjhwfjysurpx7koiti	235	9	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	235	10	78.0	78.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	235	11	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	235	12	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	236	1	4.2	4.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	236	2	Metrics	metric	NNS
work_fklnz2kdpjhwfjysurpx7koiti	236	3	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	236	4	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	5	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	236	6	thoroughly	thoroughly	RB
work_fklnz2kdpjhwfjysurpx7koiti	236	7	evaluate	evaluate	VB
work_fklnz2kdpjhwfjysurpx7koiti	236	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	236	9	performance	performance	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	236	11	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	236	12	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	236	13	ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	14	procedure	procedure	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	236	16	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	236	17	rely	rely	VBP
work_fklnz2kdpjhwfjysurpx7koiti	236	18	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	236	19	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	236	20	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	236	22	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	236	23	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	236	24	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	25	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	236	26	metrics	metric	NNS
work_fklnz2kdpjhwfjysurpx7koiti	236	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	237	1	Consider	consider	VB
work_fklnz2kdpjhwfjysurpx7koiti	237	2	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	237	3	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	237	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	237	5	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	237	6	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	237	7	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	237	8	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	237	9	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	237	10	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	237	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	237	12	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	238	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	239	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	240	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	240	2	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	240	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	240	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	240	5	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	240	6	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	240	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	240	8	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	240	9	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	240	10	–	–	:
work_fklnz2kdpjhwfjysurpx7koiti	240	11	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	240	12	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	240	13	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	240	14	standard	standard	NN
work_fklnz2kdpjhwfjysurpx7koiti	240	15	rank-	rank-	NN
work_fklnz2kdpjhwfjysurpx7koiti	240	16	ing	ing	NNP
work_fklnz2kdpjhwfjysurpx7koiti	240	17	rG(A	rG(A	NNP
work_fklnz2kdpjhwfjysurpx7koiti	240	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	240	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	240	20	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	240	21	predicted	predict	VBN
work_fklnz2kdpjhwfjysurpx7koiti	240	22	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	240	23	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	240	24	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	240	25	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	240	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	240	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	241	1	4.2.1	4.2.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	241	2	Pairwise	Pairwise	NNP
work_fklnz2kdpjhwfjysurpx7koiti	241	3	Accuracy	Accuracy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	241	4	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	241	5	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	241	6	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	241	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	241	8	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	241	9	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	241	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	241	11	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	241	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	241	13	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	241	14	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	241	15	consider	consider	VB
work_fklnz2kdpjhwfjysurpx7koiti	241	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	241	17	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	241	18	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	241	19	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	241	20	choosing	choose	VBG
work_fklnz2kdpjhwfjysurpx7koiti	241	21	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	241	22	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	241	23	three	three	CD
work_fklnz2kdpjhwfjysurpx7koiti	241	24	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	241	25	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	241	26	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	241	27	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	241	28	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	241	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	241	30	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	241	31	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	241	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	1	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	242	3	case	case	NN
work_fklnz2kdpjhwfjysurpx7koiti	242	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	5	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	6	being	be	VBG
work_fklnz2kdpjhwfjysurpx7koiti	242	7	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	242	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	242	9	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	242	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	242	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	242	12	equal	equal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	242	13	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	14	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	242	15	unknown	unknown	JJ
work_fklnz2kdpjhwfjysurpx7koiti	242	16	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	17	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	18	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	242	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	242	20	respectively	respectively	RB
work_fklnz2kdpjhwfjysurpx7koiti	242	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	242	22	com-	com-	NN
work_fklnz2kdpjhwfjysurpx7koiti	242	23	pared	pare	VBD
work_fklnz2kdpjhwfjysurpx7koiti	242	24	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	25	a2	a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	26	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	242	27	L(a1,a2	L(a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	29	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	242	30			NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	31			NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	32			NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	33	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	242	34	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	35	r(ai	r(ai	NNS
work_fklnz2kdpjhwfjysurpx7koiti	242	36	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	37	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	242	38	r(aj	r(aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	39	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	40	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	242	41	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	242	42	r(ai	r(ai	NNS
work_fklnz2kdpjhwfjysurpx7koiti	242	43	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	44	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	242	45	r(aj	r(aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	242	46	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	242	47	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	242	48	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	243	1	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	2	r(ai	r(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	243	3	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	4	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	243	5	r(aj	r(aj	LS
work_fklnz2kdpjhwfjysurpx7koiti	243	6	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	7	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	8	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	243	9	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	11	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	243	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	243	14	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	243	15	compute	compute	VBP
work_fklnz2kdpjhwfjysurpx7koiti	243	16	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	243	17	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	243	18	standard	standard	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	19	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	243	20	LG(a1,a2	lg(a1,a2	CD
work_fklnz2kdpjhwfjysurpx7koiti	243	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	22	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	243	23	predicted	predict	VBD
work_fklnz2kdpjhwfjysurpx7koiti	243	24	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	243	25	LP	LP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	243	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	27	a1,a2	a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	243	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	29	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	30	above	above	RB
work_fklnz2kdpjhwfjysurpx7koiti	243	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	243	32	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	243	33	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	243	34	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	243	35	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	36	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	37	PW(A	PW(A	.
work_fklnz2kdpjhwfjysurpx7koiti	243	38	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	243	39	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	40	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	243	41	particular	particular	JJ
work_fklnz2kdpjhwfjysurpx7koiti	243	42	ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	43	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	44	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	243	45	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	243	46	simply	simply	RB
work_fklnz2kdpjhwfjysurpx7koiti	243	47	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	243	48	fraction	fraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	243	49	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	243	50	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	243	51	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	243	52	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	243	53	correctly	correctly	RB
work_fklnz2kdpjhwfjysurpx7koiti	243	54	classified	classified	JJ
work_fklnz2kdpjhwfjysurpx7koiti	243	55	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	243	56	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	244	1	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	244	2	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	244	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	244	4	predicted	predict	VBN
work_fklnz2kdpjhwfjysurpx7koiti	244	5	label	label	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	6	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	244	7	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	244	8	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	244	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	244	10	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	244	11	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	244	12	standard	standard	JJ
work_fklnz2kdpjhwfjysurpx7koiti	244	13	label	label	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	14	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	244	15	PW(A	pw(a	LS
work_fklnz2kdpjhwfjysurpx7koiti	244	16	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	17	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	244	18	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	244	19	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	244	20	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	244	21	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	22	1{LG(ai	1{LG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	244	24	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	26	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	244	27	LP	LP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	28	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	29	ai	ai	UH
work_fklnz2kdpjhwfjysurpx7koiti	244	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	244	31	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	33	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	34	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	244	35	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	244	36	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	244	37	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	244	38	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	244	39	4.2.2	4.2.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	244	40	Ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	244	41	Correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	42	Coefficients	coefficient	NNS
work_fklnz2kdpjhwfjysurpx7koiti	244	43	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	244	44	second	second	JJ
work_fklnz2kdpjhwfjysurpx7koiti	244	45	type	type	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	46	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	244	47	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	48	assesses	assess	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	244	49	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	244	50	rank	rank	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	51	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	52	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	244	53	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	244	54	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	244	55	permutations	permutation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	244	56	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	57	gold	gold	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	58	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	244	59	standard	standard	NN
work_fklnz2kdpjhwfjysurpx7koiti	244	60	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	244	61	predicted	predict	VBD
work_fklnz2kdpjhwfjysurpx7koiti	244	62	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	244	63	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	245	1	Many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	245	2	studies	study	NNS
work_fklnz2kdpjhwfjysurpx7koiti	245	3	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	245	4	Kendall	Kendall	NNP
work_fklnz2kdpjhwfjysurpx7koiti	245	5	’s	’s	,
work_fklnz2kdpjhwfjysurpx7koiti	245	6	tau	tau	NN
work_fklnz2kdpjhwfjysurpx7koiti	245	7	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	245	8	Kendall	Kendall	NNP
work_fklnz2kdpjhwfjysurpx7koiti	245	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	245	10	1938	1938	CD
work_fklnz2kdpjhwfjysurpx7koiti	245	11	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	245	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	245	13	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	245	14	measures	measure	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	245	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	245	16	total	total	JJ
work_fklnz2kdpjhwfjysurpx7koiti	245	17	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	245	18	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	245	19	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	245	20	inversions	inversion	NNS
work_fklnz2kdpjhwfjysurpx7koiti	245	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	245	22	while	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	245	23	others	other	NNS
work_fklnz2kdpjhwfjysurpx7koiti	245	24	prefer	prefer	VBP
work_fklnz2kdpjhwfjysurpx7koiti	245	25	Spearman	Spearman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	245	26	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	245	27	rho	rho	NN
work_fklnz2kdpjhwfjysurpx7koiti	245	28	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	245	29	Spearman	Spearman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	245	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	245	31	1904	1904	CD
work_fklnz2kdpjhwfjysurpx7koiti	245	32	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	245	33	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	245	34	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	245	35	measures	measure	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	245	36	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	245	37	L1	L1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	245	38	distance	distance	NN
work_fklnz2kdpjhwfjysurpx7koiti	245	39	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	245	40	ranks	rank	NNS
work_fklnz2kdpjhwfjysurpx7koiti	245	41	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	246	1	285	285	CD
work_fklnz2kdpjhwfjysurpx7koiti	246	2	Kendall	Kendall	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	3	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	246	4	tau	tau	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	5	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	6	coefficient	coefficient	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	7	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	246	8	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	246	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	10	τb	τb	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	11	version	version	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	13	Kendall	Kendall	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	14	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	246	15	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	16	metric	metric	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	18	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	19	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	246	20	in-	in-	RB
work_fklnz2kdpjhwfjysurpx7koiti	246	21	corporates	corporate	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	246	22	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	23	correction	correction	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	24	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	25	ties	tie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	246	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	27	Kruskal	Kruskal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	29	1958	1958	CD
work_fklnz2kdpjhwfjysurpx7koiti	246	30	;	;	:
work_fklnz2kdpjhwfjysurpx7koiti	246	31	Dou	Dou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	32	et	et	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	33	al	al	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	34	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	35	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	36	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	246	37	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	38	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	246	39	τb	τb	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	40	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	246	41	P	p	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	42	−Q√	−Q√	.
work_fklnz2kdpjhwfjysurpx7koiti	246	43	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	44	P	p	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	45	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	246	46	Q	q	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	47	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	246	48	X0	X0	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	49	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	50	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	246	51	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	52	P	p	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	53	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	246	54	Q	q	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	55	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	246	56	Y0	y0	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	57	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	246	58	where	where	WRB
work_fklnz2kdpjhwfjysurpx7koiti	246	59	P	p	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	60	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	246	61	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	62	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	63	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	64	concordant	concordant	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	65	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	246	66	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	67	Q	Q	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	68	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	246	69	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	70	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	71	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	72	discordant	discordant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	246	73	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	246	74	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	75	X0	X0	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	76	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	246	77	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	78	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	79	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	80	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	246	81	tied	tie	VBN
work_fklnz2kdpjhwfjysurpx7koiti	246	82	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	83	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	84	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	246	85	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	86	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	246	87	Y0	Y0	NNP
work_fklnz2kdpjhwfjysurpx7koiti	246	88	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	246	89	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	90	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	91	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	92	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	246	93	tied	tie	VBN
work_fklnz2kdpjhwfjysurpx7koiti	246	94	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	246	95	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	246	96	second	second	JJ
work_fklnz2kdpjhwfjysurpx7koiti	246	97	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	246	98	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	247	1	Given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	247	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	3	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	247	4	rank-	rank-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	247	5	ings	ing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	247	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	247	7	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	8	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	9	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	247	10	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	13	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	247	14	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	247	15	standard	standard	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	16	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	17	rG(A	rg(a	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	247	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	21	predicted	predict	VBN
work_fklnz2kdpjhwfjysurpx7koiti	247	22	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	247	23	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	24	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	25	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	27	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	28	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	247	29	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	247	30	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	247	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	32	aj	aj	PRP
work_fklnz2kdpjhwfjysurpx7koiti	247	33	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	247	34	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	247	35	•	•	VB
work_fklnz2kdpjhwfjysurpx7koiti	247	36	concordant	concordant	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	37	iff	iff	VBP
work_fklnz2kdpjhwfjysurpx7koiti	247	38	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	39	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	247	40	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	247	41	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	42	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	247	43	strict	strict	JJ
work_fklnz2kdpjhwfjysurpx7koiti	247	44	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	247	45	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	247	46	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	247	47	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	247	48	elements	element	NNS
work_fklnz2kdpjhwfjysurpx7koiti	247	49	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	50	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	247	51	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	52	rG(ai	rG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	53	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	54	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	247	55	rG(aj	rg(aj	XX
work_fklnz2kdpjhwfjysurpx7koiti	247	56	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	57	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	247	58	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	59	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	60	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	61	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	62	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	247	63	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	64	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	65	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	66	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	67	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	247	68	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	247	69	rG(ai	rG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	70	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	71	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	247	72	rG(aj	rG(aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	73	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	74	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	247	75	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	76	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	77	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	247	78	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	79	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	247	80	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	81	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	82	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	247	83	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	247	84	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	248	1	•	•	VB
work_fklnz2kdpjhwfjysurpx7koiti	248	2	discordant	discordant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	248	3	iff	iff	VBP
work_fklnz2kdpjhwfjysurpx7koiti	248	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	248	5	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	248	6	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	248	7	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	248	8	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	248	9	inverted	inverted	JJ
work_fklnz2kdpjhwfjysurpx7koiti	248	10	strict	strict	JJ
work_fklnz2kdpjhwfjysurpx7koiti	248	11	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	248	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	248	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	248	14	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	248	15	elements	element	NNS
work_fklnz2kdpjhwfjysurpx7koiti	248	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	248	17	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	248	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	248	19	rG(ai	rG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	20	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	21	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	248	22	rG(aj	rg(aj	XX
work_fklnz2kdpjhwfjysurpx7koiti	248	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	24	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	248	25	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	26	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	27	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	248	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	29	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	248	30	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	31	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	32	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	34	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	248	35	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	248	36	rG(ai	rG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	37	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	38	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	248	39	rG(aj	rG(aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	40	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	41	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	248	42	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	43	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	44	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	45	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	46	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	248	47	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	48	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	49	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	248	50	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	248	51	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	249	1	•	•	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	2	tied	tie	VBD
work_fklnz2kdpjhwfjysurpx7koiti	249	3	iff	iff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	4	rG(ai	rG(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	5	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	6	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	249	7	rG(aj	rg(aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	249	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	9	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	249	10	rP	rP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	11	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	12	ai	ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	13	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	14	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	249	15	rP	rp	FW
work_fklnz2kdpjhwfjysurpx7koiti	249	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	17	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	249	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	249	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	250	1	Spearman	Spearman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	2	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	250	3	rho	rho	VB
work_fklnz2kdpjhwfjysurpx7koiti	250	4	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	5	coefficient	coefficient	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	6	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	7	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	250	8	n	n	JJ
work_fklnz2kdpjhwfjysurpx7koiti	250	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	250	10	sized	sized	JJ
work_fklnz2kdpjhwfjysurpx7koiti	250	11	ranked	rank	VBN
work_fklnz2kdpjhwfjysurpx7koiti	250	12	lists	list	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	13	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	14	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	15	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	16	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	250	17	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	18	yi	yi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	19	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	250	21	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	22	Spearman	Spearman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	23	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	24	coefficient	coefficient	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	25	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	250	26	defined	define	VBN
work_fklnz2kdpjhwfjysurpx7koiti	250	27	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	29	Pearson	Pearson	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	30	cor-	cor-	RB
work_fklnz2kdpjhwfjysurpx7koiti	250	31	relation	relation	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	32	coefficient	coefficient	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	33	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	34	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	35	ranks	rank	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	36	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	37	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	38	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	250	39	ρ	ρ	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	40	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	250	41	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	250	42	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	250	43	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	44	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	45	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	46	x̄	x̄	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	47	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	48	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	250	49	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	50	yi	yi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	51	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	52	ȳ	ȳ	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	53	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	54	√∑	√∑	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	55	i	i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	56	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	57	xi	xi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	58	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	59	x̄)2	x̄)2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	60	·	·	NFP
work_fklnz2kdpjhwfjysurpx7koiti	250	61	∑	∑	NFP
work_fklnz2kdpjhwfjysurpx7koiti	250	62	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	250	63	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	250	64	yi	yi	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	65	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	66	ȳ)2	ȳ)2	NN
work_fklnz2kdpjhwfjysurpx7koiti	250	67	Here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	250	68	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	250	69	x̄	x̄	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	70	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	250	71	ȳ	ȳ	NNP
work_fklnz2kdpjhwfjysurpx7koiti	250	72	denote	denote	VBP
work_fklnz2kdpjhwfjysurpx7koiti	250	73	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	74	means	mean	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	75	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	76	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	77	values	value	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	78	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	250	79	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	250	80	respective	respective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	250	81	lists	list	NNS
work_fklnz2kdpjhwfjysurpx7koiti	250	82	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	251	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	251	2	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	251	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	251	4	standard	standard	JJ
work_fklnz2kdpjhwfjysurpx7koiti	251	5	procedure	procedure	NN
work_fklnz2kdpjhwfjysurpx7koiti	251	6	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	251	7	handling	handle	VBG
work_fklnz2kdpjhwfjysurpx7koiti	251	8	ties	tie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	251	9	correctly	correctly	RB
work_fklnz2kdpjhwfjysurpx7koiti	251	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	252	1	Tied	tie	VBN
work_fklnz2kdpjhwfjysurpx7koiti	252	2	values	value	NNS
work_fklnz2kdpjhwfjysurpx7koiti	252	3	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	252	4	assigned	assign	VBN
work_fklnz2kdpjhwfjysurpx7koiti	252	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	252	6	average	average	NN
work_fklnz2kdpjhwfjysurpx7koiti	252	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	252	8	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	252	9	ranks	rank	NNS
work_fklnz2kdpjhwfjysurpx7koiti	252	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	252	11	items	item	NNS
work_fklnz2kdpjhwfjysurpx7koiti	252	12	sharing	share	VBG
work_fklnz2kdpjhwfjysurpx7koiti	252	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	252	14	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	252	15	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	252	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	252	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	252	18	ranked	rank	VBN
work_fklnz2kdpjhwfjysurpx7koiti	252	19	list	list	NN
work_fklnz2kdpjhwfjysurpx7koiti	252	20	sorted	sort	VBN
work_fklnz2kdpjhwfjysurpx7koiti	252	21	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	252	22	ascending	ascend	VBG
work_fklnz2kdpjhwfjysurpx7koiti	252	23	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	252	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	252	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	252	26	values	value	NNS
work_fklnz2kdpjhwfjysurpx7koiti	252	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	253	1	Handling	handle	VBG
work_fklnz2kdpjhwfjysurpx7koiti	253	2	Inversions	Inversions	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	253	3	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	253	4	annotating	annotate	VBG
work_fklnz2kdpjhwfjysurpx7koiti	253	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	253	6	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	253	7	some-	some-	VBP
work_fklnz2kdpjhwfjysurpx7koiti	253	8	times	time	NNS
work_fklnz2kdpjhwfjysurpx7koiti	253	9	observed	observe	VBD
work_fklnz2kdpjhwfjysurpx7koiti	253	10	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	253	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	253	12	ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	253	13	itself	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	253	14	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	253	15	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	253	16	clear	clear	JJ
work_fklnz2kdpjhwfjysurpx7koiti	253	17	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	253	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	253	19	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	253	20	disagreed	disagree	VBD
work_fklnz2kdpjhwfjysurpx7koiti	253	21	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	253	22	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	253	23	end	end	NN
work_fklnz2kdpjhwfjysurpx7koiti	253	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	253	25	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	253	26	particular	particular	JJ
work_fklnz2kdpjhwfjysurpx7koiti	253	27	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	253	28	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	253	29	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	253	30	count	count	VB
work_fklnz2kdpjhwfjysurpx7koiti	253	31	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	253	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	253	33	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	253	34	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	253	35	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	253	36	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	254	1	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	254	2	transitioning	transition	VBG
work_fklnz2kdpjhwfjysurpx7koiti	254	3	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	254	4	soft	soft	JJ
work_fklnz2kdpjhwfjysurpx7koiti	254	5	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	254	6	hard	hard	JJ
work_fklnz2kdpjhwfjysurpx7koiti	254	7	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	254	8	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	254	9	alpha	alpha	NN
work_fklnz2kdpjhwfjysurpx7koiti	254	10	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	254	11	beta	beta	NN
work_fklnz2kdpjhwfjysurpx7koiti	254	12	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	255	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	255	2	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	255	3	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	255	4	report	report	VBP
work_fklnz2kdpjhwfjysurpx7koiti	255	5	average	average	JJ
work_fklnz2kdpjhwfjysurpx7koiti	255	6	absolute	absolute	JJ
work_fklnz2kdpjhwfjysurpx7koiti	255	7	values	value	NNS
work_fklnz2kdpjhwfjysurpx7koiti	255	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	255	9	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	255	10	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	255	11	coefficients	coefficient	NNS
work_fklnz2kdpjhwfjysurpx7koiti	255	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	255	13	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	255	14	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	255	15	properly	properly	RB
work_fklnz2kdpjhwfjysurpx7koiti	255	16	ac-	ac-	RB
work_fklnz2kdpjhwfjysurpx7koiti	255	17	count	count	VB
work_fklnz2kdpjhwfjysurpx7koiti	255	18	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	255	19	anticorrelations	anticorrelation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	255	20	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	256	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	256	2	test	test	NN
work_fklnz2kdpjhwfjysurpx7koiti	256	3	set	set	VBD
work_fklnz2kdpjhwfjysurpx7koiti	256	4	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	256	5	contains	contain	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	256	6	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	256	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	256	8	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	256	9	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	256	10	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	256	11	larger	large	JJR
work_fklnz2kdpjhwfjysurpx7koiti	256	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	256	13	so	so	CC
work_fklnz2kdpjhwfjysurpx7koiti	256	14	there	there	EX
work_fklnz2kdpjhwfjysurpx7koiti	256	15	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	256	16	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	256	17	need	need	NN
work_fklnz2kdpjhwfjysurpx7koiti	256	18	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	256	19	account	account	VB
work_fklnz2kdpjhwfjysurpx7koiti	256	20	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	256	21	inversions	inversion	NNS
work_fklnz2kdpjhwfjysurpx7koiti	256	22	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	256	23	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	256	24	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	256	25	size	size	NN
work_fklnz2kdpjhwfjysurpx7koiti	256	26	2	2	CD
work_fklnz2kdpjhwfjysurpx7koiti	256	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	257	1	4.3	4.3	CD
work_fklnz2kdpjhwfjysurpx7koiti	257	2	Results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	257	3	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	257	4	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	257	5	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	257	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	257	7	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	257	8	use	use	VBP
work_fklnz2kdpjhwfjysurpx7koiti	257	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	257	10	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	257	11	metrics	metric	NNS
work_fklnz2kdpjhwfjysurpx7koiti	257	12	mentioned	mention	VBN
work_fklnz2kdpjhwfjysurpx7koiti	257	13	above	above	RB
work_fklnz2kdpjhwfjysurpx7koiti	257	14	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	257	15	compare	compare	VB
work_fklnz2kdpjhwfjysurpx7koiti	257	16	several	several	JJ
work_fklnz2kdpjhwfjysurpx7koiti	257	17	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	257	18	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	257	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	258	1	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	258	2	Baseline	Baseline	NNP
work_fklnz2kdpjhwfjysurpx7koiti	258	3	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	258	4	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	258	5	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	258	6	simply	simply	RB
work_fklnz2kdpjhwfjysurpx7koiti	258	7	reflects	reflect	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	258	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	258	9	original	original	JJ
work_fklnz2kdpjhwfjysurpx7koiti	258	10	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	258	11	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	258	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	258	13	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	258	14	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	258	15	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	258	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	259	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	259	2	classify	classify	VBP
work_fklnz2kdpjhwfjysurpx7koiti	259	3	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	4	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	5	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	259	6	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	7	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	259	8	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	259	9	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	10	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	259	11	given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	259	12	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	259	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	14	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	259	15	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	259	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	259	17	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	259	18	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	259	19	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	259	20	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	259	21	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	259	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	23	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	24	described	describe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	259	25	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	26	Section	Section	NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	27	2.1.3	2.1.3	CD
work_fklnz2kdpjhwfjysurpx7koiti	259	28	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	29	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	30	follows	follow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	259	31	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	259	32	Lbaseline(a1,a2	Lbaseline(a1,a2	NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	33	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	34	=	=	NFP
work_fklnz2kdpjhwfjysurpx7koiti	259	35			NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	36			NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	37			NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	38	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	259	39	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	40	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	41	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	259	42	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	259	43	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	44	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	259	45	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	259	46	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	259	47	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	259	48	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	49	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	259	50	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	259	51	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	259	52	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	259	53	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	259	54	=	=	NNS
work_fklnz2kdpjhwfjysurpx7koiti	259	55	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	260	1	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	2	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	260	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	260	4	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	260	5	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	260	6	=	=	SYM
work_fklnz2kdpjhwfjysurpx7koiti	260	7	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	260	8	Since	since	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	9	score(ai	score(ai	NNP
work_fklnz2kdpjhwfjysurpx7koiti	260	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	260	11	aj	aj	NNP
work_fklnz2kdpjhwfjysurpx7koiti	260	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	260	13	represents	represent	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	260	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	260	15	weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	260	16	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	260	17	strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	260	18	score	score	NN
work_fklnz2kdpjhwfjysurpx7koiti	260	19	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	260	21	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	260	22	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	260	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	260	24	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	260	25	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	260	26	positive	positive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	260	27	value	value	NN
work_fklnz2kdpjhwfjysurpx7koiti	260	28	means	mean	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	260	29	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	260	30	higher	high	JJR
work_fklnz2kdpjhwfjysurpx7koiti	260	31	likelihood	likelihood	NN
work_fklnz2kdpjhwfjysurpx7koiti	260	32	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	33	ai	ai	VB
work_fklnz2kdpjhwfjysurpx7koiti	260	34	being	be	VBG
work_fklnz2kdpjhwfjysurpx7koiti	260	35	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	260	36	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	260	37	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	260	38	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	260	39	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	40	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	260	41	left	left	NN
work_fklnz2kdpjhwfjysurpx7koiti	260	42	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	260	43	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	44	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	260	45	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	260	46	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	260	47	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	261	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	2	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	3	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	261	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	261	5	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	261	6	observe	observe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	261	7	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	261	9	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	261	10	micro	micro	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	11	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	261	12	averaged	averaged	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	13	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	261	14	pairwise	pairwise	VB
work_fklnz2kdpjhwfjysurpx7koiti	261	15	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	261	17	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	18	defined	define	VBN
work_fklnz2kdpjhwfjysurpx7koiti	261	19	earlier	early	RBR
work_fklnz2kdpjhwfjysurpx7koiti	261	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	261	21	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	22	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	261	23	origi-	origi-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	24	nal	nal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	261	25	Web	Web	NNP
work_fklnz2kdpjhwfjysurpx7koiti	261	26	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	27	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	261	28	48.2	48.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	261	29	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	261	31	while	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	261	33	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	261	34	mea-	mea-	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	35	sures	sure	NNS
work_fklnz2kdpjhwfjysurpx7koiti	261	36	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	261	37	undefined	undefined	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	38	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	39	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	261	40	individual	individual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	41	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	261	42	do	do	VBP
work_fklnz2kdpjhwfjysurpx7koiti	261	43	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	261	44	lead	lead	VB
work_fklnz2kdpjhwfjysurpx7koiti	261	45	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	261	46	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	261	47	coherent	coherent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	261	48	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	261	49	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	262	1	Divide	divide	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	2	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	262	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	262	4	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	262	5	Conquer	conquer	VB
work_fklnz2kdpjhwfjysurpx7koiti	262	6	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	7	divide	divide	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	262	9	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	262	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	262	11	conquer	conquer	VB
work_fklnz2kdpjhwfjysurpx7koiti	262	12	baseline	baseline	NNP
work_fklnz2kdpjhwfjysurpx7koiti	262	13	recursively	recursively	RB
work_fklnz2kdpjhwfjysurpx7koiti	262	14	splits	split	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	262	15	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	16	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	17	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	18	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	262	19	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	20	three	three	CD
work_fklnz2kdpjhwfjysurpx7koiti	262	21	subgroups	subgroup	NNS
work_fklnz2kdpjhwfjysurpx7koiti	262	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	262	23	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	262	24	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	26	left	left	JJ
work_fklnz2kdpjhwfjysurpx7koiti	262	27	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	28	weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	262	29	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	30	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	262	31	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	33	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	262	34	position	position	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	35	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	36	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	37	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	38	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	39	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	262	40	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	262	41	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	42	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	43	right	right	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	44	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	45	stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	262	46	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	262	47	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	262	48	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	262	49	given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	262	50	randomly	randomly	RB
work_fklnz2kdpjhwfjysurpx7koiti	262	51	chosen	choose	VBN
work_fklnz2kdpjhwfjysurpx7koiti	262	52	pivot	pivot	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	53	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	262	54	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	263	1	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	263	3	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	4	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	263	5	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	263	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	263	7	minor	minor	JJ
work_fklnz2kdpjhwfjysurpx7koiti	263	8	improve-	improve-	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	9	ment	ment	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	11	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	263	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	263	14	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	15	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	263	17	50.6	50.6	CD
work_fklnz2kdpjhwfjysurpx7koiti	263	18	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	19	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	263	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	263	21	its	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	263	22	main	main	JJ
work_fklnz2kdpjhwfjysurpx7koiti	263	23	benefit	benefit	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	24	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	263	25	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	26	one	one	CD
work_fklnz2kdpjhwfjysurpx7koiti	263	27	obtains	obtain	NNS
work_fklnz2kdpjhwfjysurpx7koiti	263	28	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	263	29	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	263	30	defined	define	VBN
work_fklnz2kdpjhwfjysurpx7koiti	263	31	inten-	inten-	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	32	sity	sity	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	33	scales	scale	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	263	34	rather	rather	RB
work_fklnz2kdpjhwfjysurpx7koiti	263	35	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	36	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	263	37	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	263	38	collection	collection	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	39	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	263	40	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	263	41	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	263	42	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	264	1	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	264	3	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	4	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	5	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	6	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	7	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	8	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	264	9	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	264	11	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	264	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	264	13	involves	involve	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	264	14	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	15	simi-	simi-	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	16	lar	lar	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	17	divide	divide	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	18	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	264	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	264	20	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	264	21	conquer	conquer	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	22	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	264	23	partitioning	partition	VBG
work_fklnz2kdpjhwfjysurpx7koiti	264	24	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	26	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	264	27	phase	phase	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	28	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	264	29	except	except	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	30	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	31	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	264	32	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	33	makes	make	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	264	34	use	use	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	35	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	36	syn-	syn-	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	37	onymy	onymy	VB
work_fklnz2kdpjhwfjysurpx7koiti	264	38	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	39	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	40	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	41	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	264	42	uses	use	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	264	43	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	44	syn-	syn-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	264	45	onyms	onyms	VBP
work_fklnz2kdpjhwfjysurpx7koiti	264	46	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	47	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	48	’s	’s	POS
work_fklnz2kdpjhwfjysurpx7koiti	264	49	synset	synset	VBD
work_fklnz2kdpjhwfjysurpx7koiti	264	50	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	51	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	52	headword	headword	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	53	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	54	neutral	neutral	JJ
work_fklnz2kdpjhwfjysurpx7koiti	264	55	pivot	pivot	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	56	elements	element	NNS
work_fklnz2kdpjhwfjysurpx7koiti	264	57	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	264	58	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	59	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	60	headword	headword	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	61	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	264	62	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	264	63	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	64	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	264	65	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	264	66	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	264	67	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	68	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	69	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	264	70	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	264	71	maximal	maximal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	264	72	unigram	unigram	JJ
work_fklnz2kdpjhwfjysurpx7koiti	264	73	frequency	frequency	NN
work_fklnz2kdpjhwfjysurpx7koiti	264	74	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	264	75	chosen	choose	VBN
work_fklnz2kdpjhwfjysurpx7koiti	264	76	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	264	77	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	265	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	265	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	265	3	second	second	JJ
work_fklnz2kdpjhwfjysurpx7koiti	265	4	phase	phase	NN
work_fklnz2kdpjhwfjysurpx7koiti	265	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	265	6	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	265	7	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	265	8	performs	perform	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	265	9	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	265	10	comparisons	comparison	NNS
work_fklnz2kdpjhwfjysurpx7koiti	265	11	within	within	IN
work_fklnz2kdpjhwfjysurpx7koiti	265	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	265	13	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	265	14	intense	intense	JJ
work_fklnz2kdpjhwfjysurpx7koiti	265	15	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	265	16	less	less	RBR
work_fklnz2kdpjhwfjysurpx7koiti	265	17	intense	intense	JJ
work_fklnz2kdpjhwfjysurpx7koiti	265	18	subgroups	subgroup	NNS
work_fklnz2kdpjhwfjysurpx7koiti	265	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	266	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	266	2	reim-	reim-	VBP
work_fklnz2kdpjhwfjysurpx7koiti	266	3	plement	plement	VBP
work_fklnz2kdpjhwfjysurpx7koiti	266	4	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	266	5	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	266	6	here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	266	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	266	8	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	266	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	266	10	Google	Google	NNP
work_fklnz2kdpjhwfjysurpx7koiti	266	11	N-	N-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	266	12	Grams	Grams	NNP
work_fklnz2kdpjhwfjysurpx7koiti	266	13	dataset	dataset	VBN
work_fklnz2kdpjhwfjysurpx7koiti	266	14	instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	266	15	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	266	16	online	online	JJ
work_fklnz2kdpjhwfjysurpx7koiti	266	17	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	266	18	search	search	NN
work_fklnz2kdpjhwfjysurpx7koiti	266	19	engine	engine	NN
work_fklnz2kdpjhwfjysurpx7koiti	266	20	hits	hit	NNS
work_fklnz2kdpjhwfjysurpx7koiti	266	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	267	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	267	2	observe	observe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	267	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	267	4	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	267	5	improvement	improvement	NN
work_fklnz2kdpjhwfjysurpx7koiti	267	6	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	267	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	267	8	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	267	9	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	267	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	267	11	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	267	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	267	13	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	267	14	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	267	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	268	1	Note	note	VB
work_fklnz2kdpjhwfjysurpx7koiti	268	2	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	268	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	268	4	286	286	CD
work_fklnz2kdpjhwfjysurpx7koiti	268	5	Method	Method	NNP
work_fklnz2kdpjhwfjysurpx7koiti	268	6	Pairwise	Pairwise	NNP
work_fklnz2kdpjhwfjysurpx7koiti	268	7	Accuracy	Accuracy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	268	8	Avg	Avg	NNP
work_fklnz2kdpjhwfjysurpx7koiti	268	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	269	1	τ	τ	LS
work_fklnz2kdpjhwfjysurpx7koiti	269	2	Avg	Avg	NNP
work_fklnz2kdpjhwfjysurpx7koiti	269	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	270	1	|τ|	|τ|	IN
work_fklnz2kdpjhwfjysurpx7koiti	270	2	Avg	Avg	NNP
work_fklnz2kdpjhwfjysurpx7koiti	270	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	271	1	ρ	ρ	IN
work_fklnz2kdpjhwfjysurpx7koiti	271	2	Avg	Avg	NNP
work_fklnz2kdpjhwfjysurpx7koiti	271	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	272	1	|ρ|	|ρ|	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	2	Web	Web	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	3	Baseline	Baseline	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	4	48.2	48.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	5	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	6	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	7	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	8	A	A	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	9	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	10	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	11	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	12	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	13	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	14	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	15	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	16	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	17	A	A	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	18	Divide	divide	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	19	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	272	20	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	272	21	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	272	22	Conquer	Conquer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	23	50.6	50.6	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	24	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	25	0.45	0.45	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	26	0.53	0.53	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	27	0.52	0.52	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	28	0.62	0.62	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	29	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	30	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	272	31	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	32	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	272	33	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	34	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	272	35	55.5	55.5	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	36	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	37	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	38	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	39	A	A	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	40	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	41	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	42	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	43	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	44	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	45	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	46	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	47	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	272	48	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	272	49	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	50	69.6	69.6	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	51	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	52	0.57	0.57	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	53	0.65	0.65	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	54	0.64	0.64	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	55	0.73	0.73	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	56	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	57	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	272	58	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	59	78.2	78.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	60	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	61	0.57	0.57	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	62	0.66	0.66	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	63	0.67	0.67	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	64	0.80	0.80	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	65	Inter	Inter	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	66	-	-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	67	Annotator	Annotator	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	68	Agreement	Agreement	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	69	78.0	78.0	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	70	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	71	0.67	0.67	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	72	0.76	0.76	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	73	0.75	0.75	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	74	0.86	0.86	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	75	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	76	3	3	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	77	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	272	78	Main	main	JJ
work_fklnz2kdpjhwfjysurpx7koiti	272	79	test	test	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	80	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	272	81	Predicted	predict	VBD
work_fklnz2kdpjhwfjysurpx7koiti	272	82	Class	Class	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	83	Weaker	Weaker	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	84	Tie	Tie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	85	Stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	272	86	True	true	JJ
work_fklnz2kdpjhwfjysurpx7koiti	272	87	Class	class	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	88	Weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	272	89	117	117	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	90	127	127	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	91	15	15	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	92	Tie	Tie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	272	93	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	94	42	42	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	95	15	15	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	96	Stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	272	97	11	11	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	98	122	122	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	99	115	115	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	100	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	101	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	272	102	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	272	103	Confusion	confusion	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	104	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	105	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	272	106	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	107	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	108	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	272	109	rank	rank	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	110	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	111	measure	measure	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	112	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	272	113	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	272	114	undefined	undefined	JJ
work_fklnz2kdpjhwfjysurpx7koiti	272	115	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	272	116	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	272	117	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	272	118	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	273	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	273	3	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	273	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	273	5	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	6	cases	case	NNS
work_fklnz2kdpjhwfjysurpx7koiti	273	7	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	273	8	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	273	9	placed	place	VBD
work_fklnz2kdpjhwfjysurpx7koiti	273	10	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	11	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	273	12	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	273	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	14	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	273	15	position	position	NN
work_fklnz2kdpjhwfjysurpx7koiti	273	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	273	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	18	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	273	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	273	20	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	273	21	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	273	22	measures	measure	NNS
work_fklnz2kdpjhwfjysurpx7koiti	273	23	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	273	24	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	273	25	handle	handle	VB
work_fklnz2kdpjhwfjysurpx7koiti	273	26	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	273	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	273	28	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	273	29	tie	tie	NN
work_fklnz2kdpjhwfjysurpx7koiti	273	30	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	273	31	corrected	correct	VBN
work_fklnz2kdpjhwfjysurpx7koiti	273	32	versions	version	NNS
work_fklnz2kdpjhwfjysurpx7koiti	273	33	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	274	1	Overall	overall	RB
work_fklnz2kdpjhwfjysurpx7koiti	274	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	274	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	274	4	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	274	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	274	6	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	274	7	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	274	8	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	274	9	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	274	10	aggregate	aggregate	VB
work_fklnz2kdpjhwfjysurpx7koiti	274	11	informa-	informa-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	274	12	tion	tion	NN
work_fklnz2kdpjhwfjysurpx7koiti	274	13	sufficiently	sufficiently	RB
work_fklnz2kdpjhwfjysurpx7koiti	274	14	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	274	15	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	274	16	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	274	17	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	274	18	level	level	NN
work_fklnz2kdpjhwfjysurpx7koiti	274	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	274	20	often	often	RB
work_fklnz2kdpjhwfjysurpx7koiti	274	21	fails	fail	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	274	22	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	274	23	make	make	VB
work_fklnz2kdpjhwfjysurpx7koiti	274	24	use	use	NN
work_fklnz2kdpjhwfjysurpx7koiti	274	25	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	274	26	transitive	transitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	274	27	inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	274	28	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	275	1	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	2	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	275	3	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	4	exploits	exploit	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	275	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	275	6	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	275	7	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	8	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	275	9	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	275	10	induce	induce	VB
work_fklnz2kdpjhwfjysurpx7koiti	275	11	significantly	significantly	RB
work_fklnz2kdpjhwfjysurpx7koiti	275	12	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	275	13	accurate	accurate	JJ
work_fklnz2kdpjhwfjysurpx7koiti	275	14	pair-	pair-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	275	15	wise	wise	JJ
work_fklnz2kdpjhwfjysurpx7koiti	275	16	labels	label	NNS
work_fklnz2kdpjhwfjysurpx7koiti	275	17	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	275	18	69.6	69.6	CD
work_fklnz2kdpjhwfjysurpx7koiti	275	19	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	20	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	275	22	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	275	23	41	41	CD
work_fklnz2kdpjhwfjysurpx7koiti	275	24	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	25	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	275	26	error	error	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	27	reduction	reduction	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	28	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	275	29	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	275	30	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	31	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	32	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	275	33	38	38	CD
work_fklnz2kdpjhwfjysurpx7koiti	275	34	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	35	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	275	36	Divide	Divide	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	37	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	275	38	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	275	39	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	275	40	Conquer	Conquer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	41	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	275	42	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	275	43	32	32	CD
work_fklnz2kdpjhwfjysurpx7koiti	275	44	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	275	45	over	over	IN
work_fklnz2kdpjhwfjysurpx7koiti	275	46	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	47	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	275	48	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	275	49	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	275	50	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	275	51	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	275	52	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	276	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	276	2	further	further	RB
work_fklnz2kdpjhwfjysurpx7koiti	276	3	see	see	VBP
work_fklnz2kdpjhwfjysurpx7koiti	276	4	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	276	5	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	276	6	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	276	7	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	276	8	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	276	9	able	able	JJ
work_fklnz2kdpjhwfjysurpx7koiti	276	10	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	276	11	exploit	exploit	VB
work_fklnz2kdpjhwfjysurpx7koiti	276	12	external	external	JJ
work_fklnz2kdpjhwfjysurpx7koiti	276	13	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	276	14	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	276	15	equiv-	equiv-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	276	16	alence	alence	NNP
work_fklnz2kdpjhwfjysurpx7koiti	276	17	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	276	18	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	276	19	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	276	20	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	276	21	synonyms	synonyms	NN
work_fklnz2kdpjhwfjysurpx7koiti	276	22	marked	mark	VBN
work_fklnz2kdpjhwfjysurpx7koiti	276	23	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	276	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	276	25	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	276	26	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	276	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	277	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	277	2	accuracy	accuracy	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	277	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	277	5	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	6	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	277	7	as	as	RB
work_fklnz2kdpjhwfjysurpx7koiti	277	8	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	277	9	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	277	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	277	11	quality	quality	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	12	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	277	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	277	14	overall	overall	JJ
work_fklnz2kdpjhwfjysurpx7koiti	277	15	ranking	ranking	JJ
work_fklnz2kdpjhwfjysurpx7koiti	277	16	increase	increase	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	17	even	even	RB
work_fklnz2kdpjhwfjysurpx7koiti	277	18	further	far	RBR
work_fklnz2kdpjhwfjysurpx7koiti	277	19	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	277	20	78.2	78.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	277	21	%	%	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	277	23	approaching	approach	VBG
work_fklnz2kdpjhwfjysurpx7koiti	277	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	277	25	human	human	JJ
work_fklnz2kdpjhwfjysurpx7koiti	277	26	inter-	inter-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	277	27	annotator	annotator	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	28	agreement	agreement	NN
work_fklnz2kdpjhwfjysurpx7koiti	277	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	278	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	2	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	4	average	average	JJ
work_fklnz2kdpjhwfjysurpx7koiti	278	5	correlation	correlation	NN
work_fklnz2kdpjhwfjysurpx7koiti	278	6	coefficients	coefficient	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	278	8	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	278	9	observe	observe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	278	10	similar	similar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	278	11	improvement	improvement	NN
work_fklnz2kdpjhwfjysurpx7koiti	278	12	trends	trend	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	13	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	278	15	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	278	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	278	17	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	278	18	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	19	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	278	20	magnitudes	magnitude	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	21	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	278	22	because	because	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	23	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	278	24	averages	average	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	25	give	give	VBP
work_fklnz2kdpjhwfjysurpx7koiti	278	26	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	278	27	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	278	29	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	278	30	weight	weight	NN
work_fklnz2kdpjhwfjysurpx7koiti	278	31	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	278	32	larger	large	JJR
work_fklnz2kdpjhwfjysurpx7koiti	278	33	ones	one	NNS
work_fklnz2kdpjhwfjysurpx7koiti	278	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	279	1	4.4	4.4	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	2	Analysis	Analysis	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	3	Confusion	Confusion	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	4	Matrices	Matrices	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	279	5	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	7	given	give	VBN
work_fklnz2kdpjhwfjysurpx7koiti	279	8	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	279	10	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	279	11	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	279	12	study	study	VB
work_fklnz2kdpjhwfjysurpx7koiti	279	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	14	confusion	confusion	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	15	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	16	obtained	obtain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	279	17	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	18	cross-	cross-	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	19	tabulating	tabulate	VBG
work_fklnz2kdpjhwfjysurpx7koiti	279	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	21	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	279	22	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	23	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	25	predicted	predict	VBN
work_fklnz2kdpjhwfjysurpx7koiti	279	26	Predicted	Predicted	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	27	Class	Class	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	28	Weaker	Weaker	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	29	Tie	Tie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	30	Stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	279	31	True	true	JJ
work_fklnz2kdpjhwfjysurpx7koiti	279	32	Class	class	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	33	Weaker	weak	JJR
work_fklnz2kdpjhwfjysurpx7koiti	279	34	177	177	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	35	29	29	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	36	53	53	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	37	Tie	tie	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	38	9	9	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	39	24	24	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	40	29	29	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	41	Stronger	strong	JJR
work_fklnz2kdpjhwfjysurpx7koiti	279	42	15	15	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	43	38	38	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	44	195	195	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	45	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	46	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	279	47	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	279	48	Confusion	confusion	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	49	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	50	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	279	51	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	279	52	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	279	53	classification	classification	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	54	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	55	every	every	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	56	unique	unique	JJ
work_fklnz2kdpjhwfjysurpx7koiti	279	57	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	58	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	59	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	279	60	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	279	61	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	279	62	ground	ground	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	63	truth	truth	NN
work_fklnz2kdpjhwfjysurpx7koiti	279	64	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	279	65	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	280	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	280	2	4	4	CD
work_fklnz2kdpjhwfjysurpx7koiti	280	3	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	280	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	280	5	confusion	confusion	NN
work_fklnz2kdpjhwfjysurpx7koiti	280	6	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	280	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	280	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	280	9	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	280	10	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	280	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	281	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	281	2	observe	observe	VBP
work_fklnz2kdpjhwfjysurpx7koiti	281	3	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	281	4	due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	281	5	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	281	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	281	7	sparsity	sparsity	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	281	9	pairwise	pairwise	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	10	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	11	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	12	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	281	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	281	15	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	16	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	281	17	predicts	predict	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	281	18	too	too	RB
work_fklnz2kdpjhwfjysurpx7koiti	281	19	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	281	20	ties	tie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	281	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	282	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	2	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	282	3	provides	provide	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	282	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	282	5	confusion	confusion	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	6	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	282	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	282	9	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	282	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	282	11	without	without	IN
work_fklnz2kdpjhwfjysurpx7koiti	282	12	external	external	JJ
work_fklnz2kdpjhwfjysurpx7koiti	282	13	equivalence	equivalence	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	14	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	15	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	282	16	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	282	17	comparison	comparison	NN
work_fklnz2kdpjhwfjysurpx7koiti	282	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	283	1	Although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	283	3	middle	middle	JJ
work_fklnz2kdpjhwfjysurpx7koiti	283	4	column	column	NN
work_fklnz2kdpjhwfjysurpx7koiti	283	5	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	283	6	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	283	7	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	283	9	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	283	10	predicts	predict	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	283	11	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	283	12	ties	tie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	283	13	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	14	humans	human	NNS
work_fklnz2kdpjhwfjysurpx7koiti	283	15	annotators	annotator	NNS
work_fklnz2kdpjhwfjysurpx7koiti	283	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	283	17	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	283	18	find	find	VBP
work_fklnz2kdpjhwfjysurpx7koiti	283	19	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	20	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	283	21	clear	clear	JJ
work_fklnz2kdpjhwfjysurpx7koiti	283	22	majority	majority	NN
work_fklnz2kdpjhwfjysurpx7koiti	283	23	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	24	all	all	DT
work_fklnz2kdpjhwfjysurpx7koiti	283	25	unique	unique	JJ
work_fklnz2kdpjhwfjysurpx7koiti	283	26	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	283	27	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	283	28	now	now	RB
work_fklnz2kdpjhwfjysurpx7koiti	283	29	correctly	correctly	RB
work_fklnz2kdpjhwfjysurpx7koiti	283	30	placed	place	VBN
work_fklnz2kdpjhwfjysurpx7koiti	283	31	along	along	IN
work_fklnz2kdpjhwfjysurpx7koiti	283	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	283	33	diagonal	diagonal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	283	34	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	284	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	284	2	confirms	confirm	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	284	3	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	284	4	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	284	5	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	284	6	successfully	successfully	RB
work_fklnz2kdpjhwfjysurpx7koiti	284	7	infers	infer	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	284	8	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	284	9	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	284	10	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	284	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	284	12	although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	284	13	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	284	14	uses	use	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	284	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	284	16	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	284	17	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	284	18	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	284	19	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	284	20	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	284	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	284	22	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	284	23	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	284	24	baseline	baseline	NN
work_fklnz2kdpjhwfjysurpx7koiti	284	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	285	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	285	2	remaining	remain	VBG
work_fklnz2kdpjhwfjysurpx7koiti	285	3	ties	tie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	285	4	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	285	5	mostly	mostly	RB
work_fklnz2kdpjhwfjysurpx7koiti	285	6	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	285	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	285	8	result	result	NN
work_fklnz2kdpjhwfjysurpx7koiti	285	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	285	10	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	285	11	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	285	12	which	which	WDT
work_fklnz2kdpjhwfjysurpx7koiti	285	13	there	there	EX
work_fklnz2kdpjhwfjysurpx7koiti	285	14	simply	simply	RB
work_fklnz2kdpjhwfjysurpx7koiti	285	15	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	285	16	no	no	DT
work_fklnz2kdpjhwfjysurpx7koiti	285	17	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	285	18	at	at	RB
work_fklnz2kdpjhwfjysurpx7koiti	285	19	all	all	RB
work_fklnz2kdpjhwfjysurpx7koiti	285	20	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	285	21	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	285	22	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	285	23	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	285	24	counts	count	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	285	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	286	1	Note	note	VB
work_fklnz2kdpjhwfjysurpx7koiti	286	2	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	286	3	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	286	4	problem	problem	NN
work_fklnz2kdpjhwfjysurpx7koiti	286	5	could	could	MD
work_fklnz2kdpjhwfjysurpx7koiti	286	6	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	286	7	instance	instance	NN
work_fklnz2kdpjhwfjysurpx7koiti	286	8	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	286	9	circum-	circum-	VBN
work_fklnz2kdpjhwfjysurpx7koiti	286	10	vented	vent	VBN
work_fklnz2kdpjhwfjysurpx7koiti	286	11	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	286	12	relying	rely	VBG
work_fklnz2kdpjhwfjysurpx7koiti	286	13	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	286	14	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	286	15	crowdsourcing	crowdsourcing	JJ
work_fklnz2kdpjhwfjysurpx7koiti	286	16	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	286	17	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	286	18	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	286	19	few	few	JJ
work_fklnz2kdpjhwfjysurpx7koiti	286	20	dispersed	disperse	VBN
work_fklnz2kdpjhwfjysurpx7koiti	286	21	tie	tie	NN
work_fklnz2kdpjhwfjysurpx7koiti	286	22	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	286	23	breakers	breaker	NNS
work_fklnz2kdpjhwfjysurpx7koiti	286	24	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	286	25	enough	enough	JJ
work_fklnz2kdpjhwfjysurpx7koiti	286	26	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	286	27	allow	allow	VB
work_fklnz2kdpjhwfjysurpx7koiti	286	28	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	286	29	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	286	30	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	286	31	correct	correct	VB
work_fklnz2kdpjhwfjysurpx7koiti	286	32	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	286	33	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	286	34	predictions	prediction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	286	35	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	287	1	Predicted	Predicted	NNP
work_fklnz2kdpjhwfjysurpx7koiti	287	2	Examples	Examples	NNP
work_fklnz2kdpjhwfjysurpx7koiti	287	3	Finally	finally	RB
work_fklnz2kdpjhwfjysurpx7koiti	287	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	287	5	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	287	6	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	287	7	6	6	CD
work_fklnz2kdpjhwfjysurpx7koiti	287	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	287	9	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	287	10	pro-	pro-	RB
work_fklnz2kdpjhwfjysurpx7koiti	287	11	vide	vide	VBP
work_fklnz2kdpjhwfjysurpx7koiti	287	12	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	287	13	selection	selection	NN
work_fklnz2kdpjhwfjysurpx7koiti	287	14	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	287	15	real	real	JJ
work_fklnz2kdpjhwfjysurpx7koiti	287	16	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	287	17	obtained	obtain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	287	18	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	287	19	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	287	20	algo-	algo-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	287	21	rithm	rithm	NNP
work_fklnz2kdpjhwfjysurpx7koiti	287	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	288	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	288	2	instance	instance	NN
work_fklnz2kdpjhwfjysurpx7koiti	288	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	288	4	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	288	5	correctly	correctly	RB
work_fklnz2kdpjhwfjysurpx7koiti	288	6	inferred	infer	VBD
work_fklnz2kdpjhwfjysurpx7koiti	288	7	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	288	8	terri-	terri-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	288	9	fying	fying	NN
work_fklnz2kdpjhwfjysurpx7koiti	288	10	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	288	11	more	more	RBR
work_fklnz2kdpjhwfjysurpx7koiti	288	12	intense	intense	JJ
work_fklnz2kdpjhwfjysurpx7koiti	288	13	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	288	14	creepy	creepy	JJ
work_fklnz2kdpjhwfjysurpx7koiti	288	15	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	288	16	scary	scary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	288	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	288	18	although	although	IN
work_fklnz2kdpjhwfjysurpx7koiti	288	19	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	288	20	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	288	21	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	288	22	counts	count	NNS
work_fklnz2kdpjhwfjysurpx7koiti	288	23	did	do	VBD
work_fklnz2kdpjhwfjysurpx7koiti	288	24	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	288	25	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	288	26	any	any	DT
work_fklnz2kdpjhwfjysurpx7koiti	288	27	explicit	explicit	JJ
work_fklnz2kdpjhwfjysurpx7koiti	288	28	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	288	29	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	288	30	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	288	31	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	288	32	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	288	33	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	289	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	2	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	289	3	cases	case	NNS
work_fklnz2kdpjhwfjysurpx7koiti	289	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	289	5	however	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	289	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	289	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	289	8	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	289	9	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	289	10	did	do	VBD
work_fklnz2kdpjhwfjysurpx7koiti	289	11	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	289	12	suffice	suffice	VB
work_fklnz2kdpjhwfjysurpx7koiti	289	13	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	289	14	draw	draw	VB
work_fklnz2kdpjhwfjysurpx7koiti	289	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	289	16	right	right	JJ
work_fklnz2kdpjhwfjysurpx7koiti	289	17	conclusions	conclusion	NNS
work_fklnz2kdpjhwfjysurpx7koiti	289	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	289	19	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	289	20	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	289	21	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	289	22	misleading	mislead	VBG
work_fklnz2kdpjhwfjysurpx7koiti	289	23	due	due	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	24	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	25	is-	is-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	289	26	sues	sue	NNS
work_fklnz2kdpjhwfjysurpx7koiti	289	27	like	like	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	28	polysemy	polysemy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	289	29	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	289	30	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	31	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	289	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	289	33	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	289	34	funny	funny	JJ
work_fklnz2kdpjhwfjysurpx7koiti	289	35	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	289	36	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	290	1	287	287	CD
work_fklnz2kdpjhwfjysurpx7koiti	290	2	Accuracy	Accuracy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	3	Prediction	Prediction	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	4	Gold	Gold	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	5	Standard	Standard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	6	Good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	7	hard	hard	RB
work_fklnz2kdpjhwfjysurpx7koiti	290	8	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	9	painful	painful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	10	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	11	hopeless	hopeless	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	12	hard	hard	RB
work_fklnz2kdpjhwfjysurpx7koiti	290	13	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	14	painful	painful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	15	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	16	hopeless	hopeless	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	17	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	18	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	19	stuffed	stuff	VBN
work_fklnz2kdpjhwfjysurpx7koiti	290	20	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	21	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	22	overflowing	overflow	VBG
work_fklnz2kdpjhwfjysurpx7koiti	290	23	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	24	overloaded	overloaded	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	25	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	26	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	27	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	28	stuffed	stuff	VBN
work_fklnz2kdpjhwfjysurpx7koiti	290	29	<	<	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	30	overflowing	overflow	VBG
work_fklnz2kdpjhwfjysurpx7koiti	290	31	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	32	overloaded	overload	VBN
work_fklnz2kdpjhwfjysurpx7koiti	290	33	unusual	unusual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	34	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	35	uncommon	uncommon	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	36	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	37	rare	rare	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	38	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	39	exceptional	exceptional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	40	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	41	extraordinary	extraordinary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	42	uncommon	uncommon	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	43	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	44	unusual	unusual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	45	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	46	rare	rare	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	47	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	48	extraordinary	extraordinary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	49	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	50	exceptional	exceptional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	51	Average	average	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	52	creepy	creepy	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	53	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	54	scary	scary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	55	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	56	sinister	sinister	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	57	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	58	frightening	frightening	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	59	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	60	terrifying	terrify	VBG
work_fklnz2kdpjhwfjysurpx7koiti	290	61	creepy	creepy	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	62	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	63	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	64	scary	scary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	65	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	66	frightening	frightening	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	67	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	68	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	69	terrifying	terrify	VBG
work_fklnz2kdpjhwfjysurpx7koiti	290	70	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	71	sinister	sinister	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	72	Bad	Bad	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	73	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	74	awake	awake	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	75	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	76	conscious	conscious	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	77	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	78	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	79	alive	alive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	80	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	81	aware	aware	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	82	alive	alive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	83	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	84	awake	awake	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	85	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	86	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	87	aware	aware	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	88	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	89	conscious	conscious	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	90	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	91	strange	strange	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	92	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	93	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	94	unusual	unusual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	95	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	96	weird	weird	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	97	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	98	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	99	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	100	funny	funny	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	101	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	102	eerie	eerie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	103	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	104	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	105	strange	strange	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	106	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	107	funny	funny	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	108	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	109	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	110	unusual	unusual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	111	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	112	weird	weird	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	113	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	290	114	eerie	eerie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	115	Table	Table	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	116	6	6	CD
work_fklnz2kdpjhwfjysurpx7koiti	290	117	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	290	118	Some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	290	119	examples	example	NNS
work_fklnz2kdpjhwfjysurpx7koiti	290	120	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	121	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	290	122	bad	bad	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	123	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	290	124	average	average	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	125	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	290	126	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	127	accu-	accu-	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	128	racy	racy	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	129	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	130	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	290	131	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	290	132	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	133	predictions	prediction	NNS
work_fklnz2kdpjhwfjysurpx7koiti	290	134	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	135	without	without	IN
work_fklnz2kdpjhwfjysurpx7koiti	290	136	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	137	infor-	infor-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	290	138	mation	mation	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	139	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	290	140	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	290	141	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	290	142	corresponding	corresponding	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	143	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	290	144	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	290	145	standard	standard	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	146	annotation	annotation	NN
work_fklnz2kdpjhwfjysurpx7koiti	290	147	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	291	1	While	while	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	2	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	291	3	show	show	VBP
work_fklnz2kdpjhwfjysurpx7koiti	291	4	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	5	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	6	gold	gold	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	7	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	291	8	standard	standard	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	9	chains	chain	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	10	here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	291	11	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	12	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	13	purposes	purpose	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	291	15	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	16	practice	practice	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	17	one	one	PRP
work_fklnz2kdpjhwfjysurpx7koiti	291	18	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	291	19	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	291	20	recombine	recombine	VB
work_fklnz2kdpjhwfjysurpx7koiti	291	21	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	291	22	[	[	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	291	23	0	0	CD
work_fklnz2kdpjhwfjysurpx7koiti	291	24	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	291	25	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	291	26	]	]	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	291	27	chains	chain	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	28	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	29	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	30	pair	pair	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	31	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	32	antonymic	antonymic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	33	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	34	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	291	35	form	form	VB
work_fklnz2kdpjhwfjysurpx7koiti	291	36	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	37	single	single	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	38	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	39	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	40	[	[	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	291	41	−1	−1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	291	42	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	291	43	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	291	44	]	]	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	291	45	that	that	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	46	visu-	visu-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	291	47	alizes	alize	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	291	48	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	49	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	50	spectrum	spectrum	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	51	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	52	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	53	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	291	54	along	along	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	55	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	56	dimension	dimension	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	57	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	291	58	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	59	adjacent	adjacent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	60	all	all	PDT
work_fklnz2kdpjhwfjysurpx7koiti	291	61	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	291	62	way	way	NN
work_fklnz2kdpjhwfjysurpx7koiti	291	63	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	291	64	removed	remove	VBN
work_fklnz2kdpjhwfjysurpx7koiti	291	65	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	291	66	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	291	67	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	68	black	black	JJ
work_fklnz2kdpjhwfjysurpx7koiti	291	69	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	291	70	glaring	glare	VBG
work_fklnz2kdpjhwfjysurpx7koiti	291	71	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	292	1	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	292	2	Extension	extension	NN
work_fklnz2kdpjhwfjysurpx7koiti	292	3	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	292	4	Multilingual	multilingual	VB
work_fklnz2kdpjhwfjysurpx7koiti	292	5	Ordering	Ordering	NNP
work_fklnz2kdpjhwfjysurpx7koiti	292	6	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	292	7	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	292	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	292	9	globally	globally	RB
work_fklnz2kdpjhwfjysurpx7koiti	292	10	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	292	11	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	292	12	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	292	13	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	292	14	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	292	15	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	292	16	easily	easily	RB
work_fklnz2kdpjhwfjysurpx7koiti	292	17	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	292	18	applied	apply	VBN
work_fklnz2kdpjhwfjysurpx7koiti	292	19	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	292	20	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	292	21	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	292	22	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	292	23	En-	en-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	292	24	glish	glish	NN
work_fklnz2kdpjhwfjysurpx7koiti	292	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	293	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	293	2	entire	entire	JJ
work_fklnz2kdpjhwfjysurpx7koiti	293	3	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	293	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	293	5	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	293	6	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	293	7	independent	independent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	293	8	as	as	RB
work_fklnz2kdpjhwfjysurpx7koiti	293	9	long	long	RB
work_fklnz2kdpjhwfjysurpx7koiti	293	10	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	293	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	293	12	required	require	VBN
work_fklnz2kdpjhwfjysurpx7koiti	293	13	resources	resource	NNS
work_fklnz2kdpjhwfjysurpx7koiti	293	14	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	293	15	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	293	16	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	293	17	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	293	18	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	293	19	number	number	NN
work_fklnz2kdpjhwfjysurpx7koiti	293	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	293	21	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	293	22	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	293	23	chosen	choose	VBN
work_fklnz2kdpjhwfjysurpx7koiti	293	24	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	294	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	294	2	morpho-	morpho-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	294	3	logically	logically	RB
work_fklnz2kdpjhwfjysurpx7koiti	294	4	rich	rich	JJ
work_fklnz2kdpjhwfjysurpx7koiti	294	5	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	294	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	294	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	294	8	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	9	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	10	step	step	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	11	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	294	12	course	course	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	13	may	may	MD
work_fklnz2kdpjhwfjysurpx7koiti	294	14	require	require	VB
work_fklnz2kdpjhwfjysurpx7koiti	294	15	additional	additional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	294	16	morphologi-	morphologi-	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	17	cal	cal	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	18	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	19	tools	tool	NNS
work_fklnz2kdpjhwfjysurpx7koiti	294	20	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	294	21	stemming	stem	VBG
work_fklnz2kdpjhwfjysurpx7koiti	294	22	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	294	23	aggregating	aggregate	VBG
work_fklnz2kdpjhwfjysurpx7koiti	294	24	fre-	fre-	NN
work_fklnz2kdpjhwfjysurpx7koiti	294	25	quencies	quencie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	294	26	across	across	IN
work_fklnz2kdpjhwfjysurpx7koiti	294	27	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	294	28	forms	form	NNS
work_fklnz2kdpjhwfjysurpx7koiti	294	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	295	1	Alternatively	alternatively	RB
work_fklnz2kdpjhwfjysurpx7koiti	295	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	295	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	295	4	cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	5	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	6	lingual	lingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	7	projection	projection	NN
work_fklnz2kdpjhwfjysurpx7koiti	295	8	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	295	9	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	295	10	possible	possible	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	11	at	at	IN
work_fklnz2kdpjhwfjysurpx7koiti	295	12	multiple	multiple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	13	levels	level	NNS
work_fklnz2kdpjhwfjysurpx7koiti	295	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	295	15	utilizing	utilize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	295	16	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	295	17	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	295	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	295	19	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	295	20	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	295	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	295	22	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	295	23	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	296	1	As	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	296	3	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	296	4	step	step	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	296	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	296	7	set	set	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	9	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	296	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	296	12	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	13	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	14	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	296	15	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	296	16	wish	wish	VBP
work_fklnz2kdpjhwfjysurpx7koiti	296	17	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	296	18	rank	rank	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	19	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	296	20	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	296	21	projected	project	VBN
work_fklnz2kdpjhwfjysurpx7koiti	296	22	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	23	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	296	24	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	296	25	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	296	26	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	296	27	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	28	necessary	necessary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	296	29	–	–	:
work_fklnz2kdpjhwfjysurpx7koiti	296	30	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	296	31	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	296	32	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	33	shown	show	VBN
work_fklnz2kdpjhwfjysurpx7koiti	296	34	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	296	35	de	de	FW
work_fklnz2kdpjhwfjysurpx7koiti	296	36	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	296	37	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	296	38	Weikum	Weikum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	296	39	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	296	40	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	296	41	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	296	42	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	297	1	Next	next	RB
work_fklnz2kdpjhwfjysurpx7koiti	297	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	297	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	297	4	outline	outline	VBP
work_fklnz2kdpjhwfjysurpx7koiti	297	5	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	297	6	projection	projection	NN
work_fklnz2kdpjhwfjysurpx7koiti	297	7	methods	method	NNS
work_fklnz2kdpjhwfjysurpx7koiti	297	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	297	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	297	10	ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	297	11	step	step	NN
work_fklnz2kdpjhwfjysurpx7koiti	297	12	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	298	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	298	2	first	first	JJ
work_fklnz2kdpjhwfjysurpx7koiti	298	3	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	298	5	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	298	6	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	298	7	pro-	pro-	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	8	jection	jection	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	9	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	298	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	298	11	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	298	12	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	13	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	298	14	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	298	15	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	298	16	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	298	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	298	18	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	298	19	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	298	21	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	298	22	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	298	23	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	298	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	298	25	same	same	JJ
work_fklnz2kdpjhwfjysurpx7koiti	298	26	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	298	27	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	298	28	described	describe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	298	29	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	298	30	Section	section	NN
work_fklnz2kdpjhwfjysurpx7koiti	298	31	2.2	2.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	298	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	299	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	299	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	299	3	second	second	JJ
work_fklnz2kdpjhwfjysurpx7koiti	299	4	method	method	NN
work_fklnz2kdpjhwfjysurpx7koiti	299	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	299	6	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	299	7	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	299	8	change	change	VBP
work_fklnz2kdpjhwfjysurpx7koiti	299	9	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	299	10	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	299	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	299	12	add	add	VB
work_fklnz2kdpjhwfjysurpx7koiti	299	13	cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	299	14	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	299	15	lingual	lingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	299	16	con-	con-	NN
work_fklnz2kdpjhwfjysurpx7koiti	299	17	straints	straint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	299	18	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	299	19	better	better	RB
work_fklnz2kdpjhwfjysurpx7koiti	299	20	inform	inform	VB
work_fklnz2kdpjhwfjysurpx7koiti	299	21	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	299	22	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	299	23	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	299	24	’s	’s	,
work_fklnz2kdpjhwfjysurpx7koiti	299	25	ad-	ad-	CC
work_fklnz2kdpjhwfjysurpx7koiti	299	26	jective	jective	NNP
work_fklnz2kdpjhwfjysurpx7koiti	299	27	ranking	ranking	NN
work_fklnz2kdpjhwfjysurpx7koiti	299	28	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	300	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	300	2	detailed	detailed	JJ
work_fklnz2kdpjhwfjysurpx7koiti	300	3	empirical	empirical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	300	4	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	300	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	300	6	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	300	7	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	300	8	remains	remain	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	300	9	future	future	JJ
work_fklnz2kdpjhwfjysurpx7koiti	300	10	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	300	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	301	1	5.1	5.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	301	2	Cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	3	-	-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	301	4	Lingual	Lingual	NNP
work_fklnz2kdpjhwfjysurpx7koiti	301	5	Pattern	Pattern	NNP
work_fklnz2kdpjhwfjysurpx7koiti	301	6	Projection	Projection	NNP
work_fklnz2kdpjhwfjysurpx7koiti	301	7	Instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	301	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	301	9	creating	create	VBG
work_fklnz2kdpjhwfjysurpx7koiti	301	10	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	11	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	301	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	301	13	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	301	14	many	many	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	15	cases	case	NNS
work_fklnz2kdpjhwfjysurpx7koiti	301	16	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	301	17	obtain	obtain	VBP
work_fklnz2kdpjhwfjysurpx7koiti	301	18	quite	quite	RB
work_fklnz2kdpjhwfjysurpx7koiti	301	19	adequate	adequate	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	20	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	301	21	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	301	22	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	301	23	us-	us-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	301	24	ing	ing	NNP
work_fklnz2kdpjhwfjysurpx7koiti	301	25	cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	26	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	27	lingual	lingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	301	28	projection	projection	NN
work_fklnz2kdpjhwfjysurpx7koiti	301	29	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	302	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	302	2	simply	simply	RB
work_fklnz2kdpjhwfjysurpx7koiti	302	3	take	take	VBP
work_fklnz2kdpjhwfjysurpx7koiti	302	4	sev-	sev-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	302	5	eral	eral	JJ
work_fklnz2kdpjhwfjysurpx7koiti	302	6	adjective	adjective	NNP
work_fklnz2kdpjhwfjysurpx7koiti	302	7	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	302	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	302	9	instantiate	instantiate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	302	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	302	11	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	302	12	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	302	13	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	302	14	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	302	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	302	16	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	302	17	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	302	18	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	302	19	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	302	20	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	302	21	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	302	22	machine	machine	NN
work_fklnz2kdpjhwfjysurpx7koiti	302	23	translation	translation	NN
work_fklnz2kdpjhwfjysurpx7koiti	302	24	system	system	NN
work_fklnz2kdpjhwfjysurpx7koiti	302	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	303	1	Filling	fill	VBG
work_fklnz2kdpjhwfjysurpx7koiti	303	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	303	3	wildcards	wildcard	NNS
work_fklnz2kdpjhwfjysurpx7koiti	303	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	303	5	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	303	6	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	303	7	tern	tern	NN
work_fklnz2kdpjhwfjysurpx7koiti	303	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	303	9	say	say	VB
work_fklnz2kdpjhwfjysurpx7koiti	303	10	‘	'	''
work_fklnz2kdpjhwfjysurpx7koiti	303	11	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	304	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	304	2	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	304	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	304	4	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	304	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	304	6	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	304	7	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	304	8	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	304	9	excellent	excellent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	304	10	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	304	11	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	304	12	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	304	13	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	304	14	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	304	15	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	304	16	excellent	excellent	JJ
work_fklnz2kdpjhwfjysurpx7koiti	304	17	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	304	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	305	1	This	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	305	2	phrase	phrase	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	305	4	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	305	5	trans-	trans-	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	6	lated	late	VBN
work_fklnz2kdpjhwfjysurpx7koiti	305	7	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	305	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	305	9	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	10	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	11	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	305	12	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	305	13	translation	translation	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	14	system	system	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	305	16	say	say	VBP
work_fklnz2kdpjhwfjysurpx7koiti	305	17	into	into	IN
work_fklnz2kdpjhwfjysurpx7koiti	305	18	German	German	NNP
work_fklnz2kdpjhwfjysurpx7koiti	305	19	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	305	20	gut	gut	NNP
work_fklnz2kdpjhwfjysurpx7koiti	305	21	aber	aber	NN
work_fklnz2kdpjhwfjysurpx7koiti	305	22	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	305	23	ausgezeich-	ausgezeich-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	305	24	net	net	NNP
work_fklnz2kdpjhwfjysurpx7koiti	305	25	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	305	26	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	306	1	Finally	finally	RB
work_fklnz2kdpjhwfjysurpx7koiti	306	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	306	3	put	put	VBN
work_fklnz2kdpjhwfjysurpx7koiti	306	4	back	back	RP
work_fklnz2kdpjhwfjysurpx7koiti	306	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	306	6	wildcards	wildcard	NNS
work_fklnz2kdpjhwfjysurpx7koiti	306	7	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	306	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	306	9	place	place	NN
work_fklnz2kdpjhwfjysurpx7koiti	306	10	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	306	11	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	306	12	translations	translation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	306	13	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	306	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	306	15	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	306	16	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	306	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	306	18	here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	306	19	gut	gut	NN
work_fklnz2kdpjhwfjysurpx7koiti	306	20	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	306	21	ausgezeichnet	ausgezeichnet	VB
work_fklnz2kdpjhwfjysurpx7koiti	306	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	306	23	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	306	24	get	get	VB
work_fklnz2kdpjhwfjysurpx7koiti	306	25	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	306	26	corresponding	corresponding	JJ
work_fklnz2kdpjhwfjysurpx7koiti	306	27	German	german	JJ
work_fklnz2kdpjhwfjysurpx7koiti	306	28	pat-	pat-	NN
work_fklnz2kdpjhwfjysurpx7koiti	306	29	tern	tern	NN
work_fklnz2kdpjhwfjysurpx7koiti	306	30	‘	'	``
work_fklnz2kdpjhwfjysurpx7koiti	306	31	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	307	1	aber	aber	NNP
work_fklnz2kdpjhwfjysurpx7koiti	307	2	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	307	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	307	4	’	'	''
work_fklnz2kdpjhwfjysurpx7koiti	307	5	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	308	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	308	2	7	7	CD
work_fklnz2kdpjhwfjysurpx7koiti	308	3	shows	show	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	308	4	various	various	JJ
work_fklnz2kdpjhwfjysurpx7koiti	308	5	German	german	JJ
work_fklnz2kdpjhwfjysurpx7koiti	308	6	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	308	7	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	308	8	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	308	9	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	308	10	obtain	obtain	VBP
work_fklnz2kdpjhwfjysurpx7koiti	308	11	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	308	12	projecting	project	VBG
work_fklnz2kdpjhwfjysurpx7koiti	308	13	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	308	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	308	15	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	308	16	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	308	17	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	308	18	described	describe	VBN
work_fklnz2kdpjhwfjysurpx7koiti	308	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	309	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	309	2	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	309	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	309	4	re-	re-	RB
work_fklnz2kdpjhwfjysurpx7koiti	309	5	peated	peate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	309	6	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	309	7	multiple	multiple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	309	8	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	309	9	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	309	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	309	11	case	case	NN
work_fklnz2kdpjhwfjysurpx7koiti	309	12	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	309	13	variants	variant	NNS
work_fklnz2kdpjhwfjysurpx7koiti	309	14	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	309	15	returned	return	VBN
work_fklnz2kdpjhwfjysurpx7koiti	309	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	309	17	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	310	1	due	due	JJ
work_fklnz2kdpjhwfjysurpx7koiti	310	2	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	310	3	morphology	morphology	NN
work_fklnz2kdpjhwfjysurpx7koiti	310	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	311	1	Most	Most	JJS
work_fklnz2kdpjhwfjysurpx7koiti	311	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	311	3	these	these	DT
work_fklnz2kdpjhwfjysurpx7koiti	311	4	translations	translation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	311	5	deliver	deliver	VBP
work_fklnz2kdpjhwfjysurpx7koiti	311	6	useful	useful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	311	7	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	311	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	312	1	Now	now	RB
work_fklnz2kdpjhwfjysurpx7koiti	312	2	that	that	IN
work_fklnz2kdpjhwfjysurpx7koiti	312	3	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	312	4	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	312	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	312	6	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	312	7	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	312	8	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	312	9	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	312	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	312	11	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	312	12	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	312	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	312	14	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	312	15	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	312	16	compute	compute	VB
work_fklnz2kdpjhwfjysurpx7koiti	312	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	312	18	pair-	pair-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	312	19	wise	wise	JJ
work_fklnz2kdpjhwfjysurpx7koiti	312	20	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	312	21	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	312	22	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	312	23	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	312	24	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	312	25	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	312	26	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	312	27	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	312	28	that	that	DT
work_fklnz2kdpjhwfjysurpx7koiti	312	29	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	312	30	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	313	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	313	2	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	313	3	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	313	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	313	5	Google	Google	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	6	n	n	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	7	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	313	8	grams	gram	NNS
work_fklnz2kdpjhwfjysurpx7koiti	313	9	cor-	cor-	IN
work_fklnz2kdpjhwfjysurpx7koiti	313	10	pora	pora	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	11	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	313	12	10	10	CD
work_fklnz2kdpjhwfjysurpx7koiti	313	13	European	european	JJ
work_fklnz2kdpjhwfjysurpx7koiti	313	14	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	313	15	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	16	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	17	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	313	18	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	19	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	313	20	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	313	21	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	22	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	313	23	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	313	24	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	313	25	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	313	26	Chinese	Chinese	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	27	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	28	LDC2010T02	ldc2010t02	CD
work_fklnz2kdpjhwfjysurpx7koiti	313	29	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	30	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	313	31	Japanese	Japanese	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	32	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	33	LDC2009T08	LDC2009T08	NNP
work_fklnz2kdpjhwfjysurpx7koiti	313	34	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	313	35	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	314	1	For	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	314	2	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	314	3	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	314	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	314	5	one	one	PRP
work_fklnz2kdpjhwfjysurpx7koiti	314	6	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	314	7	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	314	8	available	available	JJ
work_fklnz2kdpjhwfjysurpx7koiti	314	9	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	314	10	raw	raw	JJ
work_fklnz2kdpjhwfjysurpx7koiti	314	11	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	314	12	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	314	13	corpora	corpora	NN
work_fklnz2kdpjhwfjysurpx7koiti	314	14	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	314	15	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	314	16	crawling	crawling	NN
work_fklnz2kdpjhwfjysurpx7koiti	314	17	tools	tool	NNS
work_fklnz2kdpjhwfjysurpx7koiti	314	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	315	1	5.2	5.2	CD
work_fklnz2kdpjhwfjysurpx7koiti	315	2	Crosslingual	crosslingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	315	3	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	315	4	To	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	315	5	improve	improve	VB
work_fklnz2kdpjhwfjysurpx7koiti	315	6	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	315	7	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	315	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	315	9	lesser	lesser	RBR
work_fklnz2kdpjhwfjysurpx7koiti	315	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	315	11	resourced	resource	VBN
work_fklnz2kdpjhwfjysurpx7koiti	315	12	lan-	lan-	NN
work_fklnz2kdpjhwfjysurpx7koiti	315	13	guages	guage	NNS
work_fklnz2kdpjhwfjysurpx7koiti	315	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	315	15	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	315	16	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	315	17	further	further	RB
work_fklnz2kdpjhwfjysurpx7koiti	315	18	use	use	VB
work_fklnz2kdpjhwfjysurpx7koiti	315	19	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	315	20	joint	joint	JJ
work_fklnz2kdpjhwfjysurpx7koiti	315	21	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	315	22	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	315	23	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	315	24	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	315	25	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	315	26	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	315	27	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	315	28	want	want	VBP
work_fklnz2kdpjhwfjysurpx7koiti	315	29	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	315	30	transfer	transfer	VB
work_fklnz2kdpjhwfjysurpx7koiti	315	31	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	315	32	pro-	pro-	NN
work_fklnz2kdpjhwfjysurpx7koiti	315	33	cess	cess	NN
work_fklnz2kdpjhwfjysurpx7koiti	315	34	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	315	35	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	316	1	Additional	additional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	2	constraints	constraint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	316	3	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	316	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	316	5	English	English	NNP
work_fklnz2kdpjhwfjysurpx7koiti	316	6	288	288	CD
work_fklnz2kdpjhwfjysurpx7koiti	316	7	Weak	Weak	NNP
work_fklnz2kdpjhwfjysurpx7koiti	316	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	316	9	Strong	Strong	NNP
work_fklnz2kdpjhwfjysurpx7koiti	316	10	Patterns	Patterns	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	316	11	Strong	strong	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	316	13	Weak	weak	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	14	Patterns	Patterns	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	316	15	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	16	German	german	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	17	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	316	18	German	German	NNP
work_fklnz2kdpjhwfjysurpx7koiti	316	19	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	317	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	317	2	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	317	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	318	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	319	1	aber	aber	NNP
work_fklnz2kdpjhwfjysurpx7koiti	319	2	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	319	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	320	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	320	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	321	1	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	321	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	322	1	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	322	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	323	1	gerade	gerade	NNP
work_fklnz2kdpjhwfjysurpx7koiti	323	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	324	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	325	1	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	325	2	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	325	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	326	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	327	1	wenn	wenn	NNP
work_fklnz2kdpjhwfjysurpx7koiti	327	2	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	327	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	328	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	328	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	329	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	329	2	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	329	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	330	1	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	330	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	331	1	aber	aber	NNP
work_fklnz2kdpjhwfjysurpx7koiti	331	2	nur	nur	NNP
work_fklnz2kdpjhwfjysurpx7koiti	331	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	332	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	333	1	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	333	2	almost	almost	RB
work_fklnz2kdpjhwfjysurpx7koiti	333	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	334	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	335	1	und	und	RB
work_fklnz2kdpjhwfjysurpx7koiti	335	2	fast	fast	RB
work_fklnz2kdpjhwfjysurpx7koiti	335	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	336	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	336	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	337	1	though	though	IN
work_fklnz2kdpjhwfjysurpx7koiti	337	2	still	still	RB
work_fklnz2kdpjhwfjysurpx7koiti	337	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	338	1	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	338	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	339	1	aber	aber	NNP
work_fklnz2kdpjhwfjysurpx7koiti	339	2	immer	immer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	339	3	noch	noch	NNP
work_fklnz2kdpjhwfjysurpx7koiti	339	4	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	340	1	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	340	2	just	just	RB
work_fklnz2kdpjhwfjysurpx7koiti	340	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	341	1	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	341	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	342	1	nicht	nicht	NNP
work_fklnz2kdpjhwfjysurpx7koiti	342	2	nur	nur	NNP
work_fklnz2kdpjhwfjysurpx7koiti	342	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	343	1	sondern	sondern	NNP
work_fklnz2kdpjhwfjysurpx7koiti	343	2	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	344	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	345	1	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	345	2	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	345	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	346	1	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	347	1	oder	oder	NNP
work_fklnz2kdpjhwfjysurpx7koiti	347	2	sehr	sehr	NNP
work_fklnz2kdpjhwfjysurpx7koiti	347	3	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	348	1	Table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	348	2	7	7	CD
work_fklnz2kdpjhwfjysurpx7koiti	348	3	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	348	4	Examples	example	NNS
work_fklnz2kdpjhwfjysurpx7koiti	348	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	348	6	German	german	JJ
work_fklnz2kdpjhwfjysurpx7koiti	348	7	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	348	8	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	348	9	projected	project	VBN
work_fklnz2kdpjhwfjysurpx7koiti	348	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	348	11	translated	translate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	348	12	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	348	13	directly	directly	RB
work_fklnz2kdpjhwfjysurpx7koiti	348	14	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	348	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	348	16	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	348	17	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	348	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	349	1	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	349	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	349	3	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	349	4	corresponding	correspond	VBG
work_fklnz2kdpjhwfjysurpx7koiti	349	5	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	6	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	7	trans-	trans-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	349	8	lations	lation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	349	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	349	10	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	349	11	combination	combination	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	12	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	349	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	349	14	English	English	NNP
work_fklnz2kdpjhwfjysurpx7koiti	349	15	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	349	16	in-	in-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	17	formation	formation	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	18	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	349	19	allow	allow	VB
work_fklnz2kdpjhwfjysurpx7koiti	349	20	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	349	21	algorithm	algorithm	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	22	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	349	23	obtain	obtain	VB
work_fklnz2kdpjhwfjysurpx7koiti	349	24	better	well	RBR
work_fklnz2kdpjhwfjysurpx7koiti	349	25	rank-	rank-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	26	ings	ing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	349	27	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	349	28	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	349	29	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	30	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	349	31	whenever	whenever	WRB
work_fklnz2kdpjhwfjysurpx7koiti	349	32	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	349	33	non	non	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	34	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	35	English	english	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	36	target	target	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	37	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	38	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	349	39	does	do	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	349	40	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	349	41	provide	provide	VB
work_fklnz2kdpjhwfjysurpx7koiti	349	42	sufficient	sufficient	JJ
work_fklnz2kdpjhwfjysurpx7koiti	349	43	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	44	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	45	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	349	46	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	350	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	350	2	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	350	3	case	case	NN
work_fklnz2kdpjhwfjysurpx7koiti	350	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	350	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	350	6	input	input	NN
work_fklnz2kdpjhwfjysurpx7koiti	350	7	set	set	VBD
work_fklnz2kdpjhwfjysurpx7koiti	350	8	A	a	NN
work_fklnz2kdpjhwfjysurpx7koiti	350	9	contains	contain	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	350	10	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	350	11	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	350	12	multiple	multiple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	350	13	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	350	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	351	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	351	2	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	351	3	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	351	4	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	351	5	score(ai	score(ai	ADD
work_fklnz2kdpjhwfjysurpx7koiti	351	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	351	7	aj	aj	JJ
work_fklnz2kdpjhwfjysurpx7koiti	351	8	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	351	9	should	should	MD
work_fklnz2kdpjhwfjysurpx7koiti	351	10	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	351	11	set	set	VBN
work_fklnz2kdpjhwfjysurpx7koiti	351	12	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	351	13	zero	zero	CD
work_fklnz2kdpjhwfjysurpx7koiti	351	14	when	when	WRB
work_fklnz2kdpjhwfjysurpx7koiti	351	15	comparing	compare	VBG
work_fklnz2kdpjhwfjysurpx7koiti	351	16	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	351	17	across	across	IN
work_fklnz2kdpjhwfjysurpx7koiti	351	18	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	351	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	352	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	352	2	instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	352	3	link	link	VBP
work_fklnz2kdpjhwfjysurpx7koiti	352	4	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	352	5	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	352	6	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	352	7	translation	translation	NN
work_fklnz2kdpjhwfjysurpx7koiti	352	8	table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	352	9	T	t	NN
work_fklnz2kdpjhwfjysurpx7koiti	352	10	⊆	⊆	:
work_fklnz2kdpjhwfjysurpx7koiti	352	11	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	352	12	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	352	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	352	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	353	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	354	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	355	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	355	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	355	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	355	4	×	×	NN
work_fklnz2kdpjhwfjysurpx7koiti	355	5	{	{	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	355	6	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	355	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	355	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	356	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	357	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	358	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	358	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	358	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	358	4	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	358	5	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	358	6	translation	translation	NN
work_fklnz2kdpjhwfjysurpx7koiti	358	7	dictionary	dictionary	JJ
work_fklnz2kdpjhwfjysurpx7koiti	358	8	or	or	CC
work_fklnz2kdpjhwfjysurpx7koiti	358	9	phrase	phrase	NN
work_fklnz2kdpjhwfjysurpx7koiti	358	10	table	table	NN
work_fklnz2kdpjhwfjysurpx7koiti	358	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	359	1	Here	here	RB
work_fklnz2kdpjhwfjysurpx7koiti	359	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	359	3	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	359	4	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	359	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	359	6	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	359	7	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	359	8	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	359	9	T	T	NNP
work_fklnz2kdpjhwfjysurpx7koiti	359	10	signifies	signifie	NNS
work_fklnz2kdpjhwfjysurpx7koiti	359	11	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	359	12	ai	ai	VBP
work_fklnz2kdpjhwfjysurpx7koiti	359	13	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	359	14	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	359	15	translation	translation	NN
work_fklnz2kdpjhwfjysurpx7koiti	359	16	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	359	17	aj	aj	PRP
work_fklnz2kdpjhwfjysurpx7koiti	359	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	360	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	360	2	do	do	VBP
work_fklnz2kdpjhwfjysurpx7koiti	360	3	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	360	4	require	require	VB
work_fklnz2kdpjhwfjysurpx7koiti	360	5	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	360	6	bijective	bijective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	360	7	relationship	relationship	NN
work_fklnz2kdpjhwfjysurpx7koiti	360	8	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	360	9	them	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	360	10	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	360	11	i.e.	i.e.	FW
work_fklnz2kdpjhwfjysurpx7koiti	360	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	360	13	translations	translation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	360	14	need	need	VBP
work_fklnz2kdpjhwfjysurpx7koiti	360	15	n’t	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	360	16	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	360	17	unique	unique	JJ
work_fklnz2kdpjhwfjysurpx7koiti	360	18	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	360	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	361	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	361	2	objective	objective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	361	3	function	function	NN
work_fklnz2kdpjhwfjysurpx7koiti	361	4	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	361	5	augmented	augment	VBN
work_fklnz2kdpjhwfjysurpx7koiti	361	6	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	361	7	adding	add	VBG
work_fklnz2kdpjhwfjysurpx7koiti	361	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	361	9	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	361	10	term	term	NN
work_fklnz2kdpjhwfjysurpx7koiti	361	11	∑	∑	.
work_fklnz2kdpjhwfjysurpx7koiti	361	12	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	361	13	i	i	PRP
work_fklnz2kdpjhwfjysurpx7koiti	361	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	361	15	j)∈T	j)∈T	NNP
work_fklnz2kdpjhwfjysurpx7koiti	361	16	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	361	17	w′ij	w′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	361	18	+	+	SYM
work_fklnz2kdpjhwfjysurpx7koiti	361	19	s	s	NN
work_fklnz2kdpjhwfjysurpx7koiti	361	20	′	′	XX
work_fklnz2kdpjhwfjysurpx7koiti	361	21	ij)CT	ij)CT	.
work_fklnz2kdpjhwfjysurpx7koiti	361	22	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	361	23	5	5	CD
work_fklnz2kdpjhwfjysurpx7koiti	361	24	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	361	25	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	361	26	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	361	27	constant	constant	JJ
work_fklnz2kdpjhwfjysurpx7koiti	361	28	CT	CT	NNP
work_fklnz2kdpjhwfjysurpx7koiti	361	29	>	>	NN
work_fklnz2kdpjhwfjysurpx7koiti	361	30	0	0	NFP
work_fklnz2kdpjhwfjysurpx7koiti	361	31	that	that	WDT
work_fklnz2kdpjhwfjysurpx7koiti	361	32	determines	determine	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	361	33	how	how	WRB
work_fklnz2kdpjhwfjysurpx7koiti	361	34	much	much	JJ
work_fklnz2kdpjhwfjysurpx7koiti	361	35	weight	weight	NN
work_fklnz2kdpjhwfjysurpx7koiti	361	36	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	361	37	assign	assign	VBP
work_fklnz2kdpjhwfjysurpx7koiti	361	38	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	361	39	translations	translation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	361	40	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	361	41	opposed	oppose	VBN
work_fklnz2kdpjhwfjysurpx7koiti	361	42	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	361	43	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	361	44	corpus	corpus	NNP
work_fklnz2kdpjhwfjysurpx7koiti	361	45	count	count	NNP
work_fklnz2kdpjhwfjysurpx7koiti	361	46	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	361	47	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	362	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	362	2	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	362	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	362	4	extended	extend	VBN
work_fklnz2kdpjhwfjysurpx7koiti	362	5	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	362	6	adding	add	VBG
work_fklnz2kdpjhwfjysurpx7koiti	362	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	362	8	following	follow	VBG
work_fklnz2kdpjhwfjysurpx7koiti	362	9	extra	extra	JJ
work_fklnz2kdpjhwfjysurpx7koiti	362	10	constraints	constraint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	362	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	363	1	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	2	−w′ijCT	−w′ijCT	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	3	<	<	XX
work_fklnz2kdpjhwfjysurpx7koiti	363	4	−dmax	−dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	5	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	363	7	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	8	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	363	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	363	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	364	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	365	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	366	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	366	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	366	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	366	4	dij	dij	NN
work_fklnz2kdpjhwfjysurpx7koiti	366	5	+	+	CC
work_fklnz2kdpjhwfjysurpx7koiti	366	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	366	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	366	8	−w′ij)CT	−w′ij)CT	NNP
work_fklnz2kdpjhwfjysurpx7koiti	366	9	≥−dmax	≥−dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	366	10	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	366	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	366	12	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	366	13	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	366	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	366	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	367	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	368	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	369	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	369	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	369	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	369	4	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	5	+	+	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	6	s	s	NN
work_fklnz2kdpjhwfjysurpx7koiti	369	7	′	′	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	8	ijCT	ijCT	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	9	>	>	XX
work_fklnz2kdpjhwfjysurpx7koiti	369	10	dmax	dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	11	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	369	13	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	14	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	369	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	369	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	370	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	371	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	372	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	372	2	N	n	NN
work_fklnz2kdpjhwfjysurpx7koiti	372	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	372	4	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	5	−	−	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	6	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	372	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	372	8	−s′ij)CT	−s′ij)ct	NN
work_fklnz2kdpjhwfjysurpx7koiti	372	9	≤	≤	NN
work_fklnz2kdpjhwfjysurpx7koiti	372	10	dmax	dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	11	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	372	13	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	14	∈{1	∈{1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	372	15	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	372	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	373	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	374	1	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	375	1	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	2	N	N	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	3	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	375	4	w′ij	w′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	5	∈{0	∈{0	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	375	8	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	375	9	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	11	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	12	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	13	T	T	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	14	s′ij	s′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	15	∈{0	∈{0	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	17	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	375	18	}	}	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	375	19	∀i	∀i	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	20	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	21	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	22	∈	∈	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	23	T	T	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	24	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	375	25	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	26	di	di	NN
work_fklnz2kdpjhwfjysurpx7koiti	375	27	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	28	j	j	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	29	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	30	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	375	31	before	before	RB
work_fklnz2kdpjhwfjysurpx7koiti	375	32	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	33	encode	encode	NNP
work_fklnz2kdpjhwfjysurpx7koiti	375	34	distances	distance	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	35	be-	be-	XX
work_fklnz2kdpjhwfjysurpx7koiti	375	36	tween	tween	NN
work_fklnz2kdpjhwfjysurpx7koiti	375	37	positions	position	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	38	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	375	39	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	40	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	375	41	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	375	42	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	375	43	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	375	44	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	375	45	now	now	RB
work_fklnz2kdpjhwfjysurpx7koiti	375	46	also	also	RB
work_fklnz2kdpjhwfjysurpx7koiti	375	47	include	include	VBP
work_fklnz2kdpjhwfjysurpx7koiti	375	48	cross	cross	JJ
work_fklnz2kdpjhwfjysurpx7koiti	375	49	-	-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	375	50	lingual	lingual	JJ
work_fklnz2kdpjhwfjysurpx7koiti	375	51	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	52	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	375	53	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	54	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	375	55	different	different	JJ
work_fklnz2kdpjhwfjysurpx7koiti	375	56	lan-	lan-	NN
work_fklnz2kdpjhwfjysurpx7koiti	375	57	guages	guage	NNS
work_fklnz2kdpjhwfjysurpx7koiti	375	58	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	376	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	376	2	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	376	3	constraints	constraint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	376	4	encourage	encourage	VBP
work_fklnz2kdpjhwfjysurpx7koiti	376	5	translational	translational	JJ
work_fklnz2kdpjhwfjysurpx7koiti	376	6	equivalents	equivalent	NNS
work_fklnz2kdpjhwfjysurpx7koiti	376	7	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	376	8	remain	remain	VB
work_fklnz2kdpjhwfjysurpx7koiti	376	9	close	close	JJ
work_fklnz2kdpjhwfjysurpx7koiti	376	10	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	376	11	each	each	DT
work_fklnz2kdpjhwfjysurpx7koiti	376	12	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	376	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	376	14	preferably	preferably	RB
work_fklnz2kdpjhwfjysurpx7koiti	376	15	within	within	IN
work_fklnz2kdpjhwfjysurpx7koiti	376	16	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	376	17	desired	desire	VBN
work_fklnz2kdpjhwfjysurpx7koiti	376	18	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	376	19	but	but	CC
work_fklnz2kdpjhwfjysurpx7koiti	376	20	not	not	RB
work_fklnz2kdpjhwfjysurpx7koiti	376	21	strictly	strictly	RB
work_fklnz2kdpjhwfjysurpx7koiti	376	22	enforced	enforce	VBN
work_fklnz2kdpjhwfjysurpx7koiti	376	23	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	376	24	maximum	maximum	JJ
work_fklnz2kdpjhwfjysurpx7koiti	376	25	distance	distance	NN
work_fklnz2kdpjhwfjysurpx7koiti	376	26	dmax	dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	376	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	377	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	377	2	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	377	3	variables	variable	NNS
work_fklnz2kdpjhwfjysurpx7koiti	377	4	w′ij	w′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	377	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	377	6	s	s	NNP
work_fklnz2kdpjhwfjysurpx7koiti	377	7	′	′	ADD
work_fklnz2kdpjhwfjysurpx7koiti	377	8	ij	ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	377	9	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	377	10	sim-	sim-	RB
work_fklnz2kdpjhwfjysurpx7koiti	377	11	ilar	ilar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	377	12	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	377	13	wij	wij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	377	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	377	15	sij	sij	NN
work_fklnz2kdpjhwfjysurpx7koiti	377	16	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	377	17	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	377	18	standard	standard	JJ
work_fklnz2kdpjhwfjysurpx7koiti	377	19	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	377	20	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	378	1	However	however	RB
work_fklnz2kdpjhwfjysurpx7koiti	378	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	378	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	378	4	w′ij	w′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	5	become	become	VBP
work_fklnz2kdpjhwfjysurpx7koiti	378	6	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	378	7	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	378	8	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	378	9	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	378	10	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	378	11	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	12	≥−dmax	≥−dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	13	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	378	14	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	378	15	s′ij	s′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	16	become	become	VBP
work_fklnz2kdpjhwfjysurpx7koiti	378	17	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	378	18	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	378	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	378	20	only	only	RB
work_fklnz2kdpjhwfjysurpx7koiti	378	21	if	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	378	22	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	23	≤	≤	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	24	dmax	dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	378	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	379	1	If	if	IN
work_fklnz2kdpjhwfjysurpx7koiti	379	2	both	both	DT
work_fklnz2kdpjhwfjysurpx7koiti	379	3	w′ij	w′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	379	5	s′ij	s′ij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	6	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	379	7	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	379	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	379	9	then	then	RB
work_fklnz2kdpjhwfjysurpx7koiti	379	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	379	11	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	379	12	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	379	13	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	379	14	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	379	15	small	small	JJ
work_fklnz2kdpjhwfjysurpx7koiti	379	16	distance	distance	NN
work_fklnz2kdpjhwfjysurpx7koiti	379	17	−dmax	−dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	18	≤	≤	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	19	dij	dij	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	20	≤	≤	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	21	dmax	dmax	NNP
work_fklnz2kdpjhwfjysurpx7koiti	379	22	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	380	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	380	2	augmented	augment	VBN
work_fklnz2kdpjhwfjysurpx7koiti	380	3	objective	objective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	380	4	function	function	NN
work_fklnz2kdpjhwfjysurpx7koiti	380	5	explicitly	explicitly	RB
work_fklnz2kdpjhwfjysurpx7koiti	380	6	encourages	encourage	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	380	7	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	380	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	380	9	translational	translational	JJ
work_fklnz2kdpjhwfjysurpx7koiti	380	10	equivalents	equivalent	NNS
work_fklnz2kdpjhwfjysurpx7koiti	380	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	381	1	Overall	overall	RB
work_fklnz2kdpjhwfjysurpx7koiti	381	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	381	3	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	381	4	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	5	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	381	6	allows	allow	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	381	7	evi-	evi-	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	381	8	dence	dence	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	9	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	381	10	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	381	11	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	12	with	with	IN
work_fklnz2kdpjhwfjysurpx7koiti	381	13	more	more	JJR
work_fklnz2kdpjhwfjysurpx7koiti	381	14	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	15	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	16	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	381	17	improve	improve	VB
work_fklnz2kdpjhwfjysurpx7koiti	381	18	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	381	19	process	process	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	381	21	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	381	22	ordering	ordering	NN
work_fklnz2kdpjhwfjysurpx7koiti	381	23	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	381	24	lesser-	lesser-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	381	25	resourced	resourced	JJ
work_fklnz2kdpjhwfjysurpx7koiti	381	26	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	381	27	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	382	1	6	6	CD
work_fklnz2kdpjhwfjysurpx7koiti	382	2	Conclusion	Conclusion	NNP
work_fklnz2kdpjhwfjysurpx7koiti	382	3	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	4	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	382	5	work	work	NN
work_fklnz2kdpjhwfjysurpx7koiti	382	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	382	7	we	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	382	8	have	have	VBP
work_fklnz2kdpjhwfjysurpx7koiti	382	9	presented	present	VBN
work_fklnz2kdpjhwfjysurpx7koiti	382	10	an	an	DT
work_fklnz2kdpjhwfjysurpx7koiti	382	11	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	382	12	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	13	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	382	14	challenging	challenging	JJ
work_fklnz2kdpjhwfjysurpx7koiti	382	15	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	382	16	little	little	RB
work_fklnz2kdpjhwfjysurpx7koiti	382	17	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	382	18	studied	study	VBN
work_fklnz2kdpjhwfjysurpx7koiti	382	19	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	382	20	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	21	ranking	rank	VBG
work_fklnz2kdpjhwfjysurpx7koiti	382	22	words	word	NNS
work_fklnz2kdpjhwfjysurpx7koiti	382	23	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	24	terms	term	NNS
work_fklnz2kdpjhwfjysurpx7koiti	382	25	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	26	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	382	27	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	382	28	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	382	29	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	382	30	continuous	continuous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	382	31	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	382	32	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	383	1	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	383	2	address	address	VBP
work_fklnz2kdpjhwfjysurpx7koiti	383	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	383	4	issue	issue	NN
work_fklnz2kdpjhwfjysurpx7koiti	383	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	383	6	sparsity	sparsity	NN
work_fklnz2kdpjhwfjysurpx7koiti	383	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	383	8	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	383	9	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	383	10	order	order	NN
work_fklnz2kdpjhwfjysurpx7koiti	383	11	ev-	ev-	IN
work_fklnz2kdpjhwfjysurpx7koiti	383	12	idence	idence	NN
work_fklnz2kdpjhwfjysurpx7koiti	383	13	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	383	14	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	383	15	ways	way	NNS
work_fklnz2kdpjhwfjysurpx7koiti	383	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	384	1	First	first	RB
work_fklnz2kdpjhwfjysurpx7koiti	384	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	384	3	pairwise	pairwise	VB
work_fklnz2kdpjhwfjysurpx7koiti	384	4	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	384	5	scores	score	NNS
work_fklnz2kdpjhwfjysurpx7koiti	384	6	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	384	7	computed	compute	VBN
work_fklnz2kdpjhwfjysurpx7koiti	384	8	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	384	9	linguistically	linguistically	RB
work_fklnz2kdpjhwfjysurpx7koiti	384	10	intuitive	intuitive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	384	11	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	384	12	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	384	13	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	384	14	very	very	RB
work_fklnz2kdpjhwfjysurpx7koiti	384	15	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	384	16	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	384	17	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	384	18	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	384	19	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	384	20	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	384	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	385	1	Next	next	RB
work_fklnz2kdpjhwfjysurpx7koiti	385	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	385	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	385	4	Mixed	Mixed	NNP
work_fklnz2kdpjhwfjysurpx7koiti	385	5	Integer	Integer	NNP
work_fklnz2kdpjhwfjysurpx7koiti	385	6	Linear	Linear	NNP
work_fklnz2kdpjhwfjysurpx7koiti	385	7	Program	Program	NNP
work_fklnz2kdpjhwfjysurpx7koiti	385	8	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	385	9	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	385	10	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	385	11	expands	expand	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	385	12	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	385	13	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	385	14	fur-	fur-	NN
work_fklnz2kdpjhwfjysurpx7koiti	385	15	ther	ther	NN
work_fklnz2kdpjhwfjysurpx7koiti	385	16	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	385	17	inferring	infer	VBG
work_fklnz2kdpjhwfjysurpx7koiti	385	18	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	385	19	relative	relative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	385	20	relationships	relationship	NNS
work_fklnz2kdpjhwfjysurpx7koiti	385	21	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	386	1	Instead	instead	RB
work_fklnz2kdpjhwfjysurpx7koiti	386	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	386	3	making	make	VBG
work_fklnz2kdpjhwfjysurpx7koiti	386	4	ordering	order	VBG
work_fklnz2kdpjhwfjysurpx7koiti	386	5	decisions	decision	NNS
work_fklnz2kdpjhwfjysurpx7koiti	386	6	about	about	IN
work_fklnz2kdpjhwfjysurpx7koiti	386	7	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	386	8	pairs	pair	NNS
work_fklnz2kdpjhwfjysurpx7koiti	386	9	in-	in-	XX
work_fklnz2kdpjhwfjysurpx7koiti	386	10	dependently	dependently	RB
work_fklnz2kdpjhwfjysurpx7koiti	386	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	386	12	our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	386	13	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	386	14	considers	consider	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	386	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	386	16	joint	joint	JJ
work_fklnz2kdpjhwfjysurpx7koiti	386	17	decision	decision	NN
work_fklnz2kdpjhwfjysurpx7koiti	386	18	space	space	NN
work_fklnz2kdpjhwfjysurpx7koiti	386	19	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	386	20	factors	factor	NNS
work_fklnz2kdpjhwfjysurpx7koiti	386	21	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	386	22	e.g.	e.g.	RB
work_fklnz2kdpjhwfjysurpx7koiti	387	1	how	how	WRB
work_fklnz2kdpjhwfjysurpx7koiti	387	2	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	387	3	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	387	4	relate	relate	VBP
work_fklnz2kdpjhwfjysurpx7koiti	387	5	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	387	6	some	some	DT
work_fklnz2kdpjhwfjysurpx7koiti	387	7	third	third	JJ
work_fklnz2kdpjhwfjysurpx7koiti	387	8	adjective	adjective	NN
work_fklnz2kdpjhwfjysurpx7koiti	387	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	387	10	thus	thus	RB
work_fklnz2kdpjhwfjysurpx7koiti	387	11	enforcing	enforce	VBG
work_fklnz2kdpjhwfjysurpx7koiti	387	12	global	global	JJ
work_fklnz2kdpjhwfjysurpx7koiti	387	13	con-	con-	NN
work_fklnz2kdpjhwfjysurpx7koiti	387	14	straints	straint	NNS
work_fklnz2kdpjhwfjysurpx7koiti	387	15	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	387	16	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	387	17	transitivity	transitivity	NN
work_fklnz2kdpjhwfjysurpx7koiti	387	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	388	1	Our	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	388	2	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	388	3	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	388	4	general	general	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	5	enough	enough	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	6	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	388	7	allow	allow	VB
work_fklnz2kdpjhwfjysurpx7koiti	388	8	addi-	addi-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	9	tional	tional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	10	evidence	evidence	NN
work_fklnz2kdpjhwfjysurpx7koiti	388	11	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	12	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	13	synonymy	synonymy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	388	14	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	15	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	388	16	MILP	MILP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	388	17	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	388	18	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	388	19	can	can	MD
work_fklnz2kdpjhwfjysurpx7koiti	388	20	straightforwardly	straightforwardly	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	21	be	be	VB
work_fklnz2kdpjhwfjysurpx7koiti	388	22	applied	apply	VBN
work_fklnz2kdpjhwfjysurpx7koiti	388	23	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	24	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	25	word	word	NN
work_fklnz2kdpjhwfjysurpx7koiti	388	26	classes	class	NNS
work_fklnz2kdpjhwfjysurpx7koiti	388	27	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	388	28	such	such	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	29	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	30	verbs	verbs	NNP
work_fklnz2kdpjhwfjysurpx7koiti	388	31	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	388	32	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	388	33	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	388	34	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	35	other	other	JJ
work_fklnz2kdpjhwfjysurpx7koiti	388	36	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	388	37	(	(	-LRB-
work_fklnz2kdpjhwfjysurpx7koiti	388	38	monolingually	monolingually	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	39	as	as	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	40	well	well	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	41	as	as	IN
work_fklnz2kdpjhwfjysurpx7koiti	388	42	cross	cross	NN
work_fklnz2kdpjhwfjysurpx7koiti	388	43	-	-	NN
work_fklnz2kdpjhwfjysurpx7koiti	388	44	lingually	lingually	RB
work_fklnz2kdpjhwfjysurpx7koiti	388	45	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	388	46	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	389	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	389	2	overall	overall	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	3	results	result	NNS
work_fklnz2kdpjhwfjysurpx7koiti	389	4	across	across	IN
work_fklnz2kdpjhwfjysurpx7koiti	389	5	multiple	multiple	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	6	metrics	metric	NNS
work_fklnz2kdpjhwfjysurpx7koiti	389	7	are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	389	8	substan-	substan-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	9	tially	tially	RB
work_fklnz2kdpjhwfjysurpx7koiti	389	10	better	well	JJR
work_fklnz2kdpjhwfjysurpx7koiti	389	11	than	than	IN
work_fklnz2kdpjhwfjysurpx7koiti	389	12	previous	previous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	13	approaches	approach	NNS
work_fklnz2kdpjhwfjysurpx7koiti	389	14	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	389	15	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	389	16	fairly	fairly	RB
work_fklnz2kdpjhwfjysurpx7koiti	389	17	close	close	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	18	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	389	19	human	human	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	20	agreement	agreement	NN
work_fklnz2kdpjhwfjysurpx7koiti	389	21	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	389	22	this	this	DT
work_fklnz2kdpjhwfjysurpx7koiti	389	23	challenging	challenging	JJ
work_fklnz2kdpjhwfjysurpx7koiti	389	24	task	task	NN
work_fklnz2kdpjhwfjysurpx7koiti	389	25	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	390	1	Acknowledgments	acknowledgment	NNS
work_fklnz2kdpjhwfjysurpx7koiti	390	2	We	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	390	3	would	would	MD
work_fklnz2kdpjhwfjysurpx7koiti	390	4	like	like	VB
work_fklnz2kdpjhwfjysurpx7koiti	390	5	to	to	TO
work_fklnz2kdpjhwfjysurpx7koiti	390	6	thank	thank	VB
work_fklnz2kdpjhwfjysurpx7koiti	390	7	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	390	8	editor	editor	NN
work_fklnz2kdpjhwfjysurpx7koiti	390	9	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	390	10	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	390	11	anony-	anony-	NN
work_fklnz2kdpjhwfjysurpx7koiti	390	12	mous	mous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	390	13	reviewers	reviewer	NNS
work_fklnz2kdpjhwfjysurpx7koiti	390	14	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	390	15	their	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	390	16	helpful	helpful	JJ
work_fklnz2kdpjhwfjysurpx7koiti	390	17	feedback	feedback	NN
work_fklnz2kdpjhwfjysurpx7koiti	390	18	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	391	1	289	289	CD
work_fklnz2kdpjhwfjysurpx7koiti	391	2	References	reference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	391	3	Mohit	Mohit	NNP
work_fklnz2kdpjhwfjysurpx7koiti	391	4	Bansal	Bansal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	391	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	391	6	Dan	Dan	NNP
work_fklnz2kdpjhwfjysurpx7koiti	391	7	Klein	Klein	NNP
work_fklnz2kdpjhwfjysurpx7koiti	391	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	392	1	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	392	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	393	1	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	393	2	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	393	3	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	393	4	features	feature	NNS
work_fklnz2kdpjhwfjysurpx7koiti	393	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	393	6	full	full	JJ
work_fklnz2kdpjhwfjysurpx7koiti	393	7	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	393	8	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	393	9	parsing	parsing	NN
work_fklnz2kdpjhwfjysurpx7koiti	393	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	394	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	394	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	394	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	394	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	394	5	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	394	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	395	1	Thorsten	Thorsten	NNP
work_fklnz2kdpjhwfjysurpx7koiti	395	2	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	395	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	395	4	Alex	Alex	NNP
work_fklnz2kdpjhwfjysurpx7koiti	395	5	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	395	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	396	1	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	396	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	397	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	397	2	Google	Google	NNP
work_fklnz2kdpjhwfjysurpx7koiti	397	3	Web	Web	NNP
work_fklnz2kdpjhwfjysurpx7koiti	397	4	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	397	5	T	t	NN
work_fklnz2kdpjhwfjysurpx7koiti	397	6	5-gram	5-gram	CD
work_fklnz2kdpjhwfjysurpx7koiti	397	7	corpus	corpus	NN
work_fklnz2kdpjhwfjysurpx7koiti	397	8	version	version	NN
work_fklnz2kdpjhwfjysurpx7koiti	397	9	1.1	1.1	CD
work_fklnz2kdpjhwfjysurpx7koiti	397	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	398	1	LDC2006T13	LDC2006T13	NNP
work_fklnz2kdpjhwfjysurpx7koiti	398	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	399	1	Thorsten	Thorsten	NNP
work_fklnz2kdpjhwfjysurpx7koiti	399	2	Brants	Brants	NNP
work_fklnz2kdpjhwfjysurpx7koiti	399	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	399	4	Alex	Alex	NNP
work_fklnz2kdpjhwfjysurpx7koiti	399	5	Franz	Franz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	399	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	400	1	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	400	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	401	1	Web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	401	2	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	401	3	T	t	NN
work_fklnz2kdpjhwfjysurpx7koiti	401	4	5-gram	5-gram	CD
work_fklnz2kdpjhwfjysurpx7koiti	401	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	401	6	10	10	CD
work_fklnz2kdpjhwfjysurpx7koiti	401	7	European	european	JJ
work_fklnz2kdpjhwfjysurpx7koiti	401	8	languages	language	NNS
work_fklnz2kdpjhwfjysurpx7koiti	401	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	401	10	version	version	NN
work_fklnz2kdpjhwfjysurpx7koiti	401	11	1	1	CD
work_fklnz2kdpjhwfjysurpx7koiti	401	12	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	402	1	LDC2009T25	LDC2009T25	NNP
work_fklnz2kdpjhwfjysurpx7koiti	402	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	403	1	Timothy	Timothy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	403	2	Chklovski	Chklovski	NNP
work_fklnz2kdpjhwfjysurpx7koiti	403	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	403	4	Patrick	Patrick	NNP
work_fklnz2kdpjhwfjysurpx7koiti	403	5	Pantel	Pantel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	403	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	404	1	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	404	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	405	1	VerbO-	VerbO-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	405	2	cean	cean	NNP
work_fklnz2kdpjhwfjysurpx7koiti	405	3	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	405	4	Mining	mine	VBG
work_fklnz2kdpjhwfjysurpx7koiti	405	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	405	6	web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	405	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	405	8	fine	fine	RB
work_fklnz2kdpjhwfjysurpx7koiti	405	9	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	405	10	grained	grain	VBN
work_fklnz2kdpjhwfjysurpx7koiti	405	11	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	405	12	verb	verb	NN
work_fklnz2kdpjhwfjysurpx7koiti	405	13	relations	relation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	405	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	406	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	406	2	Proceedings	proceeding	NNS
work_fklnz2kdpjhwfjysurpx7koiti	406	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	406	4	EMNLP	EMNLP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	406	5	2004	2004	CD
work_fklnz2kdpjhwfjysurpx7koiti	406	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	407	1	Jacob	Jacob	NNP
work_fklnz2kdpjhwfjysurpx7koiti	407	2	Cohen	Cohen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	407	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	408	1	1960	1960	CD
work_fklnz2kdpjhwfjysurpx7koiti	408	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	409	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	409	2	coefficient	coefficient	NN
work_fklnz2kdpjhwfjysurpx7koiti	409	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	409	4	agreement	agreement	NN
work_fklnz2kdpjhwfjysurpx7koiti	409	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	409	6	nom-	nom-	DT
work_fklnz2kdpjhwfjysurpx7koiti	409	7	inal	inal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	409	8	scales	scale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	409	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	410	1	Educational	educational	JJ
work_fklnz2kdpjhwfjysurpx7koiti	410	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	410	3	Psychological	Psychological	NNP
work_fklnz2kdpjhwfjysurpx7koiti	410	4	Measure-	Measure-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	410	5	ment	ment	NN
work_fklnz2kdpjhwfjysurpx7koiti	410	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	410	7	20(1):37–46	20(1):37–46	CD
work_fklnz2kdpjhwfjysurpx7koiti	410	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	411	1	Dmitry	Dmitry	NNP
work_fklnz2kdpjhwfjysurpx7koiti	411	2	Davidov	Davidov	NNP
work_fklnz2kdpjhwfjysurpx7koiti	411	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	411	4	Ari	Ari	NNP
work_fklnz2kdpjhwfjysurpx7koiti	411	5	Rappoport	Rappoport	NNP
work_fklnz2kdpjhwfjysurpx7koiti	411	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	412	1	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	412	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	413	1	Unsuper-	Unsuper-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	413	2	vised	vised	JJ
work_fklnz2kdpjhwfjysurpx7koiti	413	3	discovery	discovery	NN
work_fklnz2kdpjhwfjysurpx7koiti	413	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	413	5	generic	generic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	413	6	relationships	relationship	NNS
work_fklnz2kdpjhwfjysurpx7koiti	413	7	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	413	8	pattern	pattern	NN
work_fklnz2kdpjhwfjysurpx7koiti	413	9	clusters	cluster	NNS
work_fklnz2kdpjhwfjysurpx7koiti	413	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	413	11	its	-PRON-	PRP$
work_fklnz2kdpjhwfjysurpx7koiti	413	12	evaluation	evaluation	NN
work_fklnz2kdpjhwfjysurpx7koiti	413	13	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	413	14	automatically	automatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	413	15	generated	generate	VBN
work_fklnz2kdpjhwfjysurpx7koiti	413	16	sat	sit	VBD
work_fklnz2kdpjhwfjysurpx7koiti	413	17	analogy	analogy	NN
work_fklnz2kdpjhwfjysurpx7koiti	413	18	questions	question	NNS
work_fklnz2kdpjhwfjysurpx7koiti	413	19	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	414	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	414	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	414	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	414	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	414	5	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	414	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	415	1	Marie	Marie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	2	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	415	3	Catherine	Catherine	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	4	de	de	IN
work_fklnz2kdpjhwfjysurpx7koiti	415	5	Marneffe	Marneffe	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	415	7	Christopher	Christopher	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	8	D.	D.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	9	Manning	Manning	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	415	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	415	12	Christopher	Christopher	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	13	Potts	Potts	NNP
work_fklnz2kdpjhwfjysurpx7koiti	415	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	416	1	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	416	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	417	1	Was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	417	2	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	417	3	good	good	JJ
work_fklnz2kdpjhwfjysurpx7koiti	417	4	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	418	1	it	-PRON-	PRP
work_fklnz2kdpjhwfjysurpx7koiti	418	2	was	be	VBD
work_fklnz2kdpjhwfjysurpx7koiti	418	3	provocative	provocative	JJ
work_fklnz2kdpjhwfjysurpx7koiti	418	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	419	1	learning	learn	VBG
work_fklnz2kdpjhwfjysurpx7koiti	419	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	419	3	meaning	meaning	NN
work_fklnz2kdpjhwfjysurpx7koiti	419	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	419	5	scalar	scalar	JJ
work_fklnz2kdpjhwfjysurpx7koiti	419	6	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	419	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	420	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	420	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	420	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	420	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	420	5	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	420	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	421	1	Gerard	Gerard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	421	2	de	de	NNP
work_fklnz2kdpjhwfjysurpx7koiti	421	3	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	421	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	421	5	Gerhard	Gerhard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	421	6	Weikum	Weikum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	421	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	422	1	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	422	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	423	1	Towards	towards	IN
work_fklnz2kdpjhwfjysurpx7koiti	423	2	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	423	3	universal	universal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	423	4	wordnet	wordnet	NN
work_fklnz2kdpjhwfjysurpx7koiti	423	5	by	by	IN
work_fklnz2kdpjhwfjysurpx7koiti	423	6	learning	learn	VBG
work_fklnz2kdpjhwfjysurpx7koiti	423	7	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	423	8	combined	combined	JJ
work_fklnz2kdpjhwfjysurpx7koiti	423	9	evi-	evi-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	423	10	dence	dence	NN
work_fklnz2kdpjhwfjysurpx7koiti	423	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	424	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	424	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	424	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	424	4	CIKM	CIKM	NNP
work_fklnz2kdpjhwfjysurpx7koiti	424	5	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	424	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	425	1	Zhicheng	Zhicheng	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	2	Dou	Dou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	425	4	Ruihua	Ruihua	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	5	Song	Song	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	425	7	Xiaojie	Xiaojie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	8	Yuan	Yuan	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	425	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	425	11	Ji	Ji	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	425	13	Rong	Rong	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	14	Wen	Wen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	425	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	426	1	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	426	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	427	1	Are	be	VBP
work_fklnz2kdpjhwfjysurpx7koiti	427	2	click	click	JJ
work_fklnz2kdpjhwfjysurpx7koiti	427	3	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	427	4	through	through	IN
work_fklnz2kdpjhwfjysurpx7koiti	427	5	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	427	6	adequate	adequate	JJ
work_fklnz2kdpjhwfjysurpx7koiti	427	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	427	8	learn-	learn-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	427	9	ing	ing	NN
work_fklnz2kdpjhwfjysurpx7koiti	427	10	web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	427	11	search	search	NN
work_fklnz2kdpjhwfjysurpx7koiti	427	12	rankings	ranking	NNS
work_fklnz2kdpjhwfjysurpx7koiti	427	13	?	?	.
work_fklnz2kdpjhwfjysurpx7koiti	428	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	428	2	Proc	Proc	NNP
work_fklnz2kdpjhwfjysurpx7koiti	428	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	429	1	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	429	2	CIKM	CIKM	NNP
work_fklnz2kdpjhwfjysurpx7koiti	429	3	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	429	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	430	1	Derek	Derek	NNP
work_fklnz2kdpjhwfjysurpx7koiti	430	2	Gross	Gross	NNP
work_fklnz2kdpjhwfjysurpx7koiti	430	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	430	4	Katherine	Katherine	NNP
work_fklnz2kdpjhwfjysurpx7koiti	430	5	J.	J.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	430	6	Miller	Miller	NNP
work_fklnz2kdpjhwfjysurpx7koiti	430	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	431	1	1990	1990	CD
work_fklnz2kdpjhwfjysurpx7koiti	431	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	432	1	Adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	432	2	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	432	3	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	432	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	433	1	International	International	NNP
work_fklnz2kdpjhwfjysurpx7koiti	433	2	Journal	Journal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	433	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	433	4	Lexicography	Lexicography	NNP
work_fklnz2kdpjhwfjysurpx7koiti	433	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	433	6	3(4):265–277	3(4):265–277	NN
work_fklnz2kdpjhwfjysurpx7koiti	433	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	434	1	Vasileios	vasileio	NNS
work_fklnz2kdpjhwfjysurpx7koiti	434	2	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	434	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	434	4	Kathleen	Kathleen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	434	5	R.	R.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	434	6	McKeown	McKeown	NNP
work_fklnz2kdpjhwfjysurpx7koiti	434	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	435	1	1993	1993	CD
work_fklnz2kdpjhwfjysurpx7koiti	435	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	436	1	Towards	towards	IN
work_fklnz2kdpjhwfjysurpx7koiti	436	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	436	3	automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	436	4	identification	identification	NN
work_fklnz2kdpjhwfjysurpx7koiti	436	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	436	6	adjecti-	adjecti-	NN
work_fklnz2kdpjhwfjysurpx7koiti	436	7	val	val	NN
work_fklnz2kdpjhwfjysurpx7koiti	436	8	scales	scale	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	436	9	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	436	10	Clustering	cluster	VBG
work_fklnz2kdpjhwfjysurpx7koiti	436	11	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	436	12	according	accord	VBG
work_fklnz2kdpjhwfjysurpx7koiti	436	13	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	436	14	meaning	meaning	NN
work_fklnz2kdpjhwfjysurpx7koiti	436	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	437	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	437	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	437	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	437	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	437	5	1993	1993	CD
work_fklnz2kdpjhwfjysurpx7koiti	437	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	438	1	Vasileios	vasileio	NNS
work_fklnz2kdpjhwfjysurpx7koiti	438	2	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	438	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	438	4	Kathleen	Kathleen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	438	5	R.	R.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	438	6	McKeown	McKeown	NNP
work_fklnz2kdpjhwfjysurpx7koiti	438	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	439	1	1997	1997	CD
work_fklnz2kdpjhwfjysurpx7koiti	439	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	440	1	Predicting	predict	VBG
work_fklnz2kdpjhwfjysurpx7koiti	440	2	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	440	3	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	440	4	orientation	orientation	NN
work_fklnz2kdpjhwfjysurpx7koiti	440	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	440	6	adjec-	adjec-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	440	7	tives	tive	NNS
work_fklnz2kdpjhwfjysurpx7koiti	440	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	441	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	441	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	441	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	441	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	441	5	1997	1997	CD
work_fklnz2kdpjhwfjysurpx7koiti	441	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	442	1	Vasileios	vasileio	NNS
work_fklnz2kdpjhwfjysurpx7koiti	442	2	Hatzivassiloglou	Hatzivassiloglou	NNP
work_fklnz2kdpjhwfjysurpx7koiti	442	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	442	4	Janyce	Janyce	NNP
work_fklnz2kdpjhwfjysurpx7koiti	442	5	M.	M.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	442	6	Wiebe	Wiebe	NNP
work_fklnz2kdpjhwfjysurpx7koiti	442	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	443	1	2000	2000	CD
work_fklnz2kdpjhwfjysurpx7koiti	443	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	444	1	Effects	effect	NNS
work_fklnz2kdpjhwfjysurpx7koiti	444	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	444	3	adjective	adjective	JJ
work_fklnz2kdpjhwfjysurpx7koiti	444	4	orientation	orientation	NN
work_fklnz2kdpjhwfjysurpx7koiti	444	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	444	6	gradability	gradability	NN
work_fklnz2kdpjhwfjysurpx7koiti	444	7	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	444	8	sen-	sen-	JJR
work_fklnz2kdpjhwfjysurpx7koiti	444	9	tence	tence	NN
work_fklnz2kdpjhwfjysurpx7koiti	444	10	subjectivity	subjectivity	NN
work_fklnz2kdpjhwfjysurpx7koiti	444	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	445	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	445	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	445	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	445	4	COLING	COLING	NNP
work_fklnz2kdpjhwfjysurpx7koiti	445	5	2000	2000	CD
work_fklnz2kdpjhwfjysurpx7koiti	445	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	446	1	Marti	Marti	NNP
work_fklnz2kdpjhwfjysurpx7koiti	446	2	Hearst	Hearst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	446	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	447	1	1992	1992	CD
work_fklnz2kdpjhwfjysurpx7koiti	447	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	448	1	Automatic	automatic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	448	2	acquisition	acquisition	NN
work_fklnz2kdpjhwfjysurpx7koiti	448	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	448	4	hyponyms	hyponyms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	448	5	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	448	6	large	large	JJ
work_fklnz2kdpjhwfjysurpx7koiti	448	7	text	text	NN
work_fklnz2kdpjhwfjysurpx7koiti	448	8	corpora	corpora	NN
work_fklnz2kdpjhwfjysurpx7koiti	448	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	449	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	449	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	449	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	449	4	COLING	COLING	NNP
work_fklnz2kdpjhwfjysurpx7koiti	449	5	1992	1992	CD
work_fklnz2kdpjhwfjysurpx7koiti	449	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	450	1	Diana	Diana	NNP
work_fklnz2kdpjhwfjysurpx7koiti	450	2	Inkpen	Inkpen	NNP
work_fklnz2kdpjhwfjysurpx7koiti	450	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	450	4	Graeme	Graeme	NNP
work_fklnz2kdpjhwfjysurpx7koiti	450	5	Hirst	Hirst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	450	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	451	1	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	451	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	452	1	Building	building	NN
work_fklnz2kdpjhwfjysurpx7koiti	452	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	452	3	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	452	4	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	452	5	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	452	6	knowledge	knowledge	NN
work_fklnz2kdpjhwfjysurpx7koiti	452	7	base	base	NN
work_fklnz2kdpjhwfjysurpx7koiti	452	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	452	9	near	near	IN
work_fklnz2kdpjhwfjysurpx7koiti	452	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	452	11	synonym	synonym	NN
work_fklnz2kdpjhwfjysurpx7koiti	452	12	dif-	dif-	CC
work_fklnz2kdpjhwfjysurpx7koiti	452	13	ferences	ference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	452	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	453	1	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	453	2	Linguistics	Linguistics	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	453	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	453	4	32(2):223–262	32(2):223–262	CD
work_fklnz2kdpjhwfjysurpx7koiti	453	5	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	454	1	Maurice	Maurice	NNP
work_fklnz2kdpjhwfjysurpx7koiti	454	2	G.	G.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	454	3	Kendall	Kendall	NNP
work_fklnz2kdpjhwfjysurpx7koiti	454	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	455	1	1938	1938	CD
work_fklnz2kdpjhwfjysurpx7koiti	455	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	456	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	456	2	new	new	JJ
work_fklnz2kdpjhwfjysurpx7koiti	456	3	measure	measure	NN
work_fklnz2kdpjhwfjysurpx7koiti	456	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	456	5	rank	rank	NN
work_fklnz2kdpjhwfjysurpx7koiti	456	6	cor-	cor-	NN
work_fklnz2kdpjhwfjysurpx7koiti	456	7	relation	relation	NN
work_fklnz2kdpjhwfjysurpx7koiti	456	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	457	1	Biometrika	Biometrika	NNP
work_fklnz2kdpjhwfjysurpx7koiti	457	2	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	457	3	30(1/2):81–93	30(1/2):81–93	CD
work_fklnz2kdpjhwfjysurpx7koiti	457	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	458	1	Adam	Adam	NNP
work_fklnz2kdpjhwfjysurpx7koiti	458	2	Kilgarriff	Kilgarriff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	458	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	459	1	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	459	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	460	1	Googleology	googleology	NN
work_fklnz2kdpjhwfjysurpx7koiti	460	2	is	be	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	460	3	bad	bad	JJ
work_fklnz2kdpjhwfjysurpx7koiti	460	4	science	science	NN
work_fklnz2kdpjhwfjysurpx7koiti	460	5	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	461	1	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	461	2	Linguistics	Linguistics	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	461	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	461	4	33(1	33(1	CD
work_fklnz2kdpjhwfjysurpx7koiti	461	5	)	)	-RRB-
work_fklnz2kdpjhwfjysurpx7koiti	461	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	462	1	William	William	NNP
work_fklnz2kdpjhwfjysurpx7koiti	462	2	H.	H.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	462	3	Kruskal	Kruskal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	462	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	463	1	1958	1958	CD
work_fklnz2kdpjhwfjysurpx7koiti	463	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	464	1	Ordinal	ordinal	JJ
work_fklnz2kdpjhwfjysurpx7koiti	464	2	measures	measure	NNS
work_fklnz2kdpjhwfjysurpx7koiti	464	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	464	4	associa-	associa-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	464	5	tion	tion	NN
work_fklnz2kdpjhwfjysurpx7koiti	464	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	465	1	Journal	Journal	NNP
work_fklnz2kdpjhwfjysurpx7koiti	465	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	465	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	465	4	American	American	NNP
work_fklnz2kdpjhwfjysurpx7koiti	465	5	Statistical	Statistical	NNP
work_fklnz2kdpjhwfjysurpx7koiti	465	6	Association	Association	NNP
work_fklnz2kdpjhwfjysurpx7koiti	465	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	465	8	53(284):814–861	53(284):814–861	CD
work_fklnz2kdpjhwfjysurpx7koiti	465	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	466	1	Jingjing	Jingjing	NNP
work_fklnz2kdpjhwfjysurpx7koiti	466	2	Liu	Liu	NNP
work_fklnz2kdpjhwfjysurpx7koiti	466	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	466	4	Stephanie	Stephanie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	466	5	Seneff	Seneff	NNP
work_fklnz2kdpjhwfjysurpx7koiti	466	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	467	1	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	467	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	467	3	Review	review	NN
work_fklnz2kdpjhwfjysurpx7koiti	467	4	senti-	senti-	NN
work_fklnz2kdpjhwfjysurpx7koiti	467	5	ment	ment	JJ
work_fklnz2kdpjhwfjysurpx7koiti	467	6	scoring	scoring	NN
work_fklnz2kdpjhwfjysurpx7koiti	467	7	via	via	IN
work_fklnz2kdpjhwfjysurpx7koiti	467	8	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	467	9	parse	parse	JJ
work_fklnz2kdpjhwfjysurpx7koiti	467	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	467	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	467	12	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	467	13	paraphrase	paraphrase	NN
work_fklnz2kdpjhwfjysurpx7koiti	467	14	paradigm	paradigm	NN
work_fklnz2kdpjhwfjysurpx7koiti	467	15	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	468	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	468	2	Proceedings	proceeding	NNS
work_fklnz2kdpjhwfjysurpx7koiti	468	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	468	4	EMNLP	EMNLP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	468	5	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	468	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	469	1	George	George	NNP
work_fklnz2kdpjhwfjysurpx7koiti	469	2	A.	A.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	469	3	Miller	Miller	NNP
work_fklnz2kdpjhwfjysurpx7koiti	469	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	470	1	1995	1995	CD
work_fklnz2kdpjhwfjysurpx7koiti	470	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	471	1	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	471	2	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	471	3	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	471	4	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	471	5	database	database	NN
work_fklnz2kdpjhwfjysurpx7koiti	471	6	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	471	7	english	english	NNP
work_fklnz2kdpjhwfjysurpx7koiti	471	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	472	1	Communications	communication	NNS
work_fklnz2kdpjhwfjysurpx7koiti	472	2	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	472	3	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	472	4	ACM	ACM	NNP
work_fklnz2kdpjhwfjysurpx7koiti	472	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	472	6	38(11):39–41	38(11):39–41	CD
work_fklnz2kdpjhwfjysurpx7koiti	472	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	473	1	Said	Said	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	2	M.	M.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	3	Mohammad	Mohammad	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	473	5	Bonnie	Bonnie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	6	J.	J.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	7	Dorr	Dorr	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	473	9	Graeme	Graeme	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	10	Hirst	Hirst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	11	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	473	12	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	473	13	Peter	Peter	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	14	D.	D.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	15	Turney	Turney	NNP
work_fklnz2kdpjhwfjysurpx7koiti	473	16	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	474	1	2013	2013	CD
work_fklnz2kdpjhwfjysurpx7koiti	474	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	475	1	Computing	compute	VBG
work_fklnz2kdpjhwfjysurpx7koiti	475	2	lexical	lexical	JJ
work_fklnz2kdpjhwfjysurpx7koiti	475	3	contrast	contrast	NN
work_fklnz2kdpjhwfjysurpx7koiti	475	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	476	1	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	476	2	Linguistics	Linguistics	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	476	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	477	1	Bo	Bo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	477	2	Pang	Pang	NNP
work_fklnz2kdpjhwfjysurpx7koiti	477	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	477	4	Lillian	Lillian	NNP
work_fklnz2kdpjhwfjysurpx7koiti	477	5	Lee	Lee	NNP
work_fklnz2kdpjhwfjysurpx7koiti	477	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	478	1	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	478	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	479	1	Opinion	opinion	NN
work_fklnz2kdpjhwfjysurpx7koiti	479	2	mining	mining	NN
work_fklnz2kdpjhwfjysurpx7koiti	479	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	479	4	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	479	5	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	479	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	480	1	Foundations	foundation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	480	2	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	480	3	Trends	Trends	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	480	4	in	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	480	5	Infor-	Infor-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	480	6	mation	mation	NN
work_fklnz2kdpjhwfjysurpx7koiti	480	7	Retrieval	Retrieval	NNP
work_fklnz2kdpjhwfjysurpx7koiti	480	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	480	9	2(1	2(1	NNP
work_fklnz2kdpjhwfjysurpx7koiti	480	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	480	11	2):1–135	2):1–135	NNP
work_fklnz2kdpjhwfjysurpx7koiti	480	12	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	480	13	January	January	NNP
work_fklnz2kdpjhwfjysurpx7koiti	480	14	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	481	1	Peter	Peter	NNP
work_fklnz2kdpjhwfjysurpx7koiti	481	2	F.	F.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	481	3	Schulam	Schulam	NNP
work_fklnz2kdpjhwfjysurpx7koiti	481	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	481	5	Christiane	Christiane	NNP
work_fklnz2kdpjhwfjysurpx7koiti	481	6	Fellbaum	Fellbaum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	481	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	482	1	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	482	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	483	1	Au-	au-	DT
work_fklnz2kdpjhwfjysurpx7koiti	483	2	tomatically	tomatically	RB
work_fklnz2kdpjhwfjysurpx7koiti	483	3	determining	determine	VBG
work_fklnz2kdpjhwfjysurpx7koiti	483	4	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	483	5	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	483	6	gradation	gradation	NN
work_fklnz2kdpjhwfjysurpx7koiti	483	7	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	483	8	ger-	ger-	JJ
work_fklnz2kdpjhwfjysurpx7koiti	483	9	man	man	NN
work_fklnz2kdpjhwfjysurpx7koiti	483	10	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	483	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	484	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	484	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	484	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	484	4	KONVENS	KONVENS	NNP
work_fklnz2kdpjhwfjysurpx7koiti	484	5	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	484	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	485	1	Vera	Vera	NNP
work_fklnz2kdpjhwfjysurpx7koiti	485	2	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	485	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	485	4	Takenobu	Takenobu	NNP
work_fklnz2kdpjhwfjysurpx7koiti	485	5	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	485	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	486	1	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	486	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	487	1	AdjS-	AdjS-	NFP
work_fklnz2kdpjhwfjysurpx7koiti	487	2	cales	cale	NNS
work_fklnz2kdpjhwfjysurpx7koiti	487	3	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	487	4	Visualizing	visualize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	487	5	differences	difference	NNS
work_fklnz2kdpjhwfjysurpx7koiti	487	6	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	487	7	adjectives	adjective	NNS
work_fklnz2kdpjhwfjysurpx7koiti	487	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	487	9	language	language	NN
work_fklnz2kdpjhwfjysurpx7koiti	487	10	learners	learner	NNS
work_fklnz2kdpjhwfjysurpx7koiti	487	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	488	1	IEICE	ieice	NN
work_fklnz2kdpjhwfjysurpx7koiti	488	2	Transactions	Transactions	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	488	3	on	on	IN
work_fklnz2kdpjhwfjysurpx7koiti	488	4	Information	Information	NNP
work_fklnz2kdpjhwfjysurpx7koiti	488	5	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	488	6	Systems	Systems	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	488	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	488	8	92(8):1542–1550	92(8):1542–1550	CD
work_fklnz2kdpjhwfjysurpx7koiti	488	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	489	1	Vera	Vera	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	2	Sheinman	Sheinman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	489	4	Takenobu	Takenobu	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	5	Tokunaga	Tokunaga	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	489	7	I.	I.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	8	Julien	Julien	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	489	10	P.	P.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	11	Schu-	Schu-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	12	lam	lam	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	13	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	489	14	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	489	15	C.	C.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	16	Fellbaum	Fellbaum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	489	17	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	490	1	2012	2012	CD
work_fklnz2kdpjhwfjysurpx7koiti	490	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	491	1	Refining	Refining	NNP
work_fklnz2kdpjhwfjysurpx7koiti	491	2	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	491	3	adjec-	adjec-	NN
work_fklnz2kdpjhwfjysurpx7koiti	491	4	tive	tive	JJ
work_fklnz2kdpjhwfjysurpx7koiti	491	5	dumbbells	dumbbell	NNS
work_fklnz2kdpjhwfjysurpx7koiti	491	6	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	491	7	intensity	intensity	NN
work_fklnz2kdpjhwfjysurpx7koiti	491	8	relations	relation	NNS
work_fklnz2kdpjhwfjysurpx7koiti	491	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	492	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	492	2	Proceed-	Proceed-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	492	3	ings	ing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	492	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	492	5	Global	Global	NNP
work_fklnz2kdpjhwfjysurpx7koiti	492	6	WordNet	WordNet	NNP
work_fklnz2kdpjhwfjysurpx7koiti	492	7	Conference	Conference	NNP
work_fklnz2kdpjhwfjysurpx7koiti	492	8	2012	2012	CD
work_fklnz2kdpjhwfjysurpx7koiti	492	9	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	493	1	Rion	Rion	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	2	Snow	Snow	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	493	4	Daniel	Daniel	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	5	Jurafsky	Jurafsky	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	493	7	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	493	8	Andrew	Andrew	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	9	Y.	Y.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	10	Ng	Ng	NNP
work_fklnz2kdpjhwfjysurpx7koiti	493	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	494	1	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	494	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	495	1	Semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	495	2	taxonomy	taxonomy	NN
work_fklnz2kdpjhwfjysurpx7koiti	495	3	induction	induction	NN
work_fklnz2kdpjhwfjysurpx7koiti	495	4	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	495	5	heterogenous	heterogenous	JJ
work_fklnz2kdpjhwfjysurpx7koiti	495	6	evi-	evi-	NNS
work_fklnz2kdpjhwfjysurpx7koiti	495	7	dence	dence	NN
work_fklnz2kdpjhwfjysurpx7koiti	495	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	496	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	496	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	496	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	496	4	COLING	COLING	NNP
work_fklnz2kdpjhwfjysurpx7koiti	496	5	/	/	SYM
work_fklnz2kdpjhwfjysurpx7koiti	496	6	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	496	7	2006	2006	CD
work_fklnz2kdpjhwfjysurpx7koiti	496	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	497	1	Charles	Charles	NNP
work_fklnz2kdpjhwfjysurpx7koiti	497	2	Spearman	Spearman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	497	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	498	1	1904	1904	CD
work_fklnz2kdpjhwfjysurpx7koiti	498	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	499	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	499	2	proof	proof	NN
work_fklnz2kdpjhwfjysurpx7koiti	499	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	499	4	measurement	measurement	NN
work_fklnz2kdpjhwfjysurpx7koiti	499	5	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	499	6	association	association	NNP
work_fklnz2kdpjhwfjysurpx7koiti	499	7	between	between	IN
work_fklnz2kdpjhwfjysurpx7koiti	499	8	two	two	CD
work_fklnz2kdpjhwfjysurpx7koiti	499	9	things	thing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	499	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	500	1	The	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	500	2	American	american	JJ
work_fklnz2kdpjhwfjysurpx7koiti	500	3	journal	journal	NN
work_fklnz2kdpjhwfjysurpx7koiti	500	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	500	5	psychology	psychology	NN
work_fklnz2kdpjhwfjysurpx7koiti	500	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	500	7	15(1):72–101	15(1):72–101	CD
work_fklnz2kdpjhwfjysurpx7koiti	500	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	501	1	Fabian	Fabian	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	2	M.	M.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	3	Suchanek	Suchanek	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	4	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	501	5	Mauro	Mauro	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	6	Sozio	Sozio	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	7	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	501	8	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	501	9	Gerhard	Gerhard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	10	Weikum	Weikum	NNP
work_fklnz2kdpjhwfjysurpx7koiti	501	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	502	1	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	502	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	503	1	SOFIE	sofie	RB
work_fklnz2kdpjhwfjysurpx7koiti	503	2	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	503	3	a	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	503	4	self	self	NN
work_fklnz2kdpjhwfjysurpx7koiti	503	5	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	503	6	organizing	organize	VBG
work_fklnz2kdpjhwfjysurpx7koiti	503	7	framework	framework	NN
work_fklnz2kdpjhwfjysurpx7koiti	503	8	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	503	9	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	503	10	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	503	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	504	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	504	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	504	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	504	4	WWW	WWW	NNP
work_fklnz2kdpjhwfjysurpx7koiti	504	5	2009	2009	CD
work_fklnz2kdpjhwfjysurpx7koiti	504	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	505	1	Maite	Maite	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	2	Taboada	Taboada	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	505	4	Julian	Julian	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	5	Brooke	Brooke	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	505	7	Milan	Milan	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	8	Tofiloskiy	Tofiloskiy	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	9	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	505	10	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	505	11	Kimberly	Kimberly	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	12	Vollz	Vollz	NNP
work_fklnz2kdpjhwfjysurpx7koiti	505	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	506	1	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	506	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	507	1	Lexicon	Lexicon	NNP
work_fklnz2kdpjhwfjysurpx7koiti	507	2	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	507	3	based	base	VBN
work_fklnz2kdpjhwfjysurpx7koiti	507	4	methods	method	NNS
work_fklnz2kdpjhwfjysurpx7koiti	507	5	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	507	6	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	507	7	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	507	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	508	1	Computational	Computational	NNP
work_fklnz2kdpjhwfjysurpx7koiti	508	2	Linguistics	Linguistics	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	508	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	509	1	Niket	Niket	NNP
work_fklnz2kdpjhwfjysurpx7koiti	509	2	Tandon	Tandon	NNP
work_fklnz2kdpjhwfjysurpx7koiti	509	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	509	4	Gerard	Gerard	NNP
work_fklnz2kdpjhwfjysurpx7koiti	509	5	de	de	NNP
work_fklnz2kdpjhwfjysurpx7koiti	509	6	Melo	Melo	NNP
work_fklnz2kdpjhwfjysurpx7koiti	509	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	510	1	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	510	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	511	1	Information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	511	2	extraction	extraction	NN
work_fklnz2kdpjhwfjysurpx7koiti	511	3	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	511	4	web	web	NN
work_fklnz2kdpjhwfjysurpx7koiti	511	5	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	511	6	scale	scale	NN
work_fklnz2kdpjhwfjysurpx7koiti	511	7	n	n	JJ
work_fklnz2kdpjhwfjysurpx7koiti	511	8	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	511	9	gram	gram	NN
work_fklnz2kdpjhwfjysurpx7koiti	511	10	data	datum	NNS
work_fklnz2kdpjhwfjysurpx7koiti	511	11	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	512	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	512	2	Proceed-	Proceed-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	512	3	ings	ing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	512	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	512	5	the	the	DT
work_fklnz2kdpjhwfjysurpx7koiti	512	6	SIGIR	SIGIR	NNP
work_fklnz2kdpjhwfjysurpx7koiti	512	7	2010	2010	CD
work_fklnz2kdpjhwfjysurpx7koiti	512	8	Web	Web	NNP
work_fklnz2kdpjhwfjysurpx7koiti	512	9	N	N	NNP
work_fklnz2kdpjhwfjysurpx7koiti	512	10	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	512	11	gram	gram	NN
work_fklnz2kdpjhwfjysurpx7koiti	512	12	Workshop	Workshop	NNP
work_fklnz2kdpjhwfjysurpx7koiti	512	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	513	1	Peter	Peter	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	2	D.	D.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	3	Turney	Turney	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	513	5	Michael	Michael	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	6	L.	L.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	7	Littman	Littman	NNP
work_fklnz2kdpjhwfjysurpx7koiti	513	8	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	514	1	2003	2003	CD
work_fklnz2kdpjhwfjysurpx7koiti	514	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	515	1	Mea-	Mea-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	515	2	suring	sure	VBG
work_fklnz2kdpjhwfjysurpx7koiti	515	3	praise	praise	NN
work_fklnz2kdpjhwfjysurpx7koiti	515	4	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	515	5	criticism	criticism	NN
work_fklnz2kdpjhwfjysurpx7koiti	515	6	:	:	:
work_fklnz2kdpjhwfjysurpx7koiti	515	7	Inference	inference	NN
work_fklnz2kdpjhwfjysurpx7koiti	515	8	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	515	9	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	515	10	orientation	orientation	NN
work_fklnz2kdpjhwfjysurpx7koiti	515	11	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	515	12	association	association	NN
work_fklnz2kdpjhwfjysurpx7koiti	515	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	516	1	ACM	ACM	NNP
work_fklnz2kdpjhwfjysurpx7koiti	516	2	Trans	Trans	NNPS
work_fklnz2kdpjhwfjysurpx7koiti	516	3	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	517	1	Inf	inf	NN
work_fklnz2kdpjhwfjysurpx7koiti	517	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	518	1	Syst	Syst	NNP
work_fklnz2kdpjhwfjysurpx7koiti	518	2	.	.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	518	3	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	518	4	21(4):315–346	21(4):315–346	CD
work_fklnz2kdpjhwfjysurpx7koiti	518	5	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	518	6	October	October	NNP
work_fklnz2kdpjhwfjysurpx7koiti	518	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	519	1	Peter	Peter	NNP
work_fklnz2kdpjhwfjysurpx7koiti	519	2	D.	D.	NNP
work_fklnz2kdpjhwfjysurpx7koiti	519	3	Turney	Turney	NNP
work_fklnz2kdpjhwfjysurpx7koiti	519	4	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	520	1	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	520	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	521	1	A	a	DT
work_fklnz2kdpjhwfjysurpx7koiti	521	2	uniform	uniform	JJ
work_fklnz2kdpjhwfjysurpx7koiti	521	3	approach	approach	NN
work_fklnz2kdpjhwfjysurpx7koiti	521	4	to	to	IN
work_fklnz2kdpjhwfjysurpx7koiti	521	5	analogies	analogy	NNS
work_fklnz2kdpjhwfjysurpx7koiti	521	6	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	521	7	synonyms	synonyms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	521	8	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	521	9	antonyms	antonyms	NNP
work_fklnz2kdpjhwfjysurpx7koiti	521	10	,	,	,
work_fklnz2kdpjhwfjysurpx7koiti	521	11	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	521	12	associations	association	NNS
work_fklnz2kdpjhwfjysurpx7koiti	521	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	522	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	522	2	Proceed-	Proceed-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	522	3	ings	ing	NNS
work_fklnz2kdpjhwfjysurpx7koiti	522	4	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	522	5	COLING	COLING	NNP
work_fklnz2kdpjhwfjysurpx7koiti	522	6	2008	2008	CD
work_fklnz2kdpjhwfjysurpx7koiti	522	7	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	523	1	Xiaofeng	Xiaofeng	NNP
work_fklnz2kdpjhwfjysurpx7koiti	523	2	Yang	Yang	NNP
work_fklnz2kdpjhwfjysurpx7koiti	523	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	523	4	Jian	Jian	NNP
work_fklnz2kdpjhwfjysurpx7koiti	523	5	Su	Su	NNP
work_fklnz2kdpjhwfjysurpx7koiti	523	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	524	1	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	524	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	525	1	Coreference	coreference	NN
work_fklnz2kdpjhwfjysurpx7koiti	525	2	resolu-	resolu-	VBZ
work_fklnz2kdpjhwfjysurpx7koiti	525	3	tion	tion	NN
work_fklnz2kdpjhwfjysurpx7koiti	525	4	using	use	VBG
work_fklnz2kdpjhwfjysurpx7koiti	525	5	semantic	semantic	JJ
work_fklnz2kdpjhwfjysurpx7koiti	525	6	relatedness	relatedness	NN
work_fklnz2kdpjhwfjysurpx7koiti	525	7	information	information	NN
work_fklnz2kdpjhwfjysurpx7koiti	525	8	from	from	IN
work_fklnz2kdpjhwfjysurpx7koiti	525	9	auto-	auto-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	525	10	matically	matically	RB
work_fklnz2kdpjhwfjysurpx7koiti	525	11	discovered	discovered	JJ
work_fklnz2kdpjhwfjysurpx7koiti	525	12	patterns	pattern	NNS
work_fklnz2kdpjhwfjysurpx7koiti	525	13	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	526	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	526	2	Proceedings	Proceedings	NNP
work_fklnz2kdpjhwfjysurpx7koiti	526	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	526	4	ACL	ACL	NNP
work_fklnz2kdpjhwfjysurpx7koiti	526	5	2007	2007	CD
work_fklnz2kdpjhwfjysurpx7koiti	526	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	527	1	Ainur	Ainur	NNP
work_fklnz2kdpjhwfjysurpx7koiti	527	2	Yessenalina	Yessenalina	NNP
work_fklnz2kdpjhwfjysurpx7koiti	527	3	and	and	CC
work_fklnz2kdpjhwfjysurpx7koiti	527	4	Claire	Claire	NNP
work_fklnz2kdpjhwfjysurpx7koiti	527	5	Cardie	Cardie	NNP
work_fklnz2kdpjhwfjysurpx7koiti	527	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	528	1	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	528	2	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	529	1	Composi-	Composi-	NNP
work_fklnz2kdpjhwfjysurpx7koiti	529	2	tional	tional	JJ
work_fklnz2kdpjhwfjysurpx7koiti	529	3	matrix	matrix	NN
work_fklnz2kdpjhwfjysurpx7koiti	529	4	-	-	HYPH
work_fklnz2kdpjhwfjysurpx7koiti	529	5	space	space	NN
work_fklnz2kdpjhwfjysurpx7koiti	529	6	models	model	NNS
work_fklnz2kdpjhwfjysurpx7koiti	529	7	for	for	IN
work_fklnz2kdpjhwfjysurpx7koiti	529	8	sentiment	sentiment	NN
work_fklnz2kdpjhwfjysurpx7koiti	529	9	analysis	analysis	NN
work_fklnz2kdpjhwfjysurpx7koiti	529	10	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	530	1	In	in	IN
work_fklnz2kdpjhwfjysurpx7koiti	530	2	Proceedings	proceeding	NNS
work_fklnz2kdpjhwfjysurpx7koiti	530	3	of	of	IN
work_fklnz2kdpjhwfjysurpx7koiti	530	4	EMNLP	EMNLP	NNP
work_fklnz2kdpjhwfjysurpx7koiti	530	5	2011	2011	CD
work_fklnz2kdpjhwfjysurpx7koiti	530	6	.	.	.
work_fklnz2kdpjhwfjysurpx7koiti	531	1	290	290	CD