id	sid	tid	token	lemma	pos
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	1	Document	document	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	2	downloaded	download	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	3	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	4	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	5	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	6	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	7	must	must	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	8	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	9	cited	cite	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	10	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	11	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	12	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	13	final	final	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	14	publication	publication	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	15	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	16	available	available	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	17	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	18	Copyright	copyright	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	19	http://dx.doi.org/10.1016/j.eswa.2012.01.140	http://dx.doi.org/10.1016/j.eswa.2012.01.140	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	20	http://hdl.handle.net/10251/34812	http://hdl.handle.net/10251/34812	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	21	Elsevier	Elsevier	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	22	Domenech	Domenech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	24	J.	J.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	25	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	26	De	De	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	27	La	La	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	28	Ossa	Ossa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	29	Perez	Perez	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	30	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	31	BA	BA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	1	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	1	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	2	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	3	Borrás	Borrás	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	5	J.	J.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	6	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	7	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	8	Salinas	Salinas	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	10	JA	JA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	2	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	3	1	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	3	2	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	3	3	Sanjuan	Sanjuan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	3	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	3	5	A.	a.	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	4	1	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	4	2	2012	2012	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	4	3	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	4	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	2	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	4	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	5	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	6	1	Expert	expert	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	6	2	Systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	6	3	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	6	4	Applications	Applications	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	6	5	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	1	39(9):8496	39(9):8496	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	2	-	-	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	3	8502	8502	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	5	doi:10.1016	doi:10.1016	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	6	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	7	j.eswa.2012.01.140	j.eswa.2012.01.140	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	7	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	2	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	4	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	7	Josep	Josep	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	8	Domenech	Domenech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	10	Bernardo	Bernardo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	11	de	de	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	12	la	la	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	13	Ossa	Ossa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	15	Julio	Julio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	16	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	18	Jose	Jose	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	19	A.	A.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	20	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	22	Ana	Ana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	23	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	24	Universitat	Universitat	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	25	Politecnica	Politecnica	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	26	de	de	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	27	Valencia	Valencia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	8	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	1	Camı́	Camı́	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	2	de	de	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	3	Vera	Vera	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	5	s	s	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	6	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	7	n	n	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	9	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	1	46022	46022	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	2	Valencia	Valencia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	3	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	4	Spain	Spain	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	5	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	6	jdomenech@upvnet.upv.es	jdomenech@upvnet.upv.es	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	7	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	8	berospe@doctor.upv.es	berospe@doctor.upv.es	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	9	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	10	{	{	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	11	jsahuqui,jagil,apont}@disca.upv.es	jsahuqui,jagil,apont}@disca.upv.es	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	12	November	November	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	13	10	10	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	15	2011	2011	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	16	Abstract	Abstract	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	17	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	18	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	19	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	20	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	21	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	22	attractive	attractive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	23	solution	solution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	24	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	25	reduce	reduce	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	27	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	28	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	29	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	30	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	10	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	2	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	4	driven	drive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	5	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	7	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	8	en-	en-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	9	gine	gine	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	10	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	11	algorithm	algorithm	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	12	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	13	guesses	guess	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	14	following	follow	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	15	actions	action	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	17	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	18	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	11	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	2	large	large	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	3	amount	amount	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	7	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	8	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	9	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	10	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	12	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	13	pre-	pre-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	14	fetching	fetching	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	15	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	16	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	17	published	publish	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	19	although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	20	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	21	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	22	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	23	over	over	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	25	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	26	two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	27	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	28	three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	29	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	30	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	31	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	32	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	33	begun	begin	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	34	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	35	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	36	successfully	successfully	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	37	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	38	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	39	com-	com-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	40	mercial	mercial	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	41	products	product	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	12	42	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	3	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	4	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	5	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	7	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	8	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	11	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	12	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	14	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	15	use	use	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	16	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	17	wide	wide	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	18	variety	variety	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	20	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	21	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	22	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	13	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	2	affects	affect	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	3	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	4	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	6	data	data	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	7	system	system	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	9	computational	computational	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	10	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	12	ac-	ac-	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	13	curacy	curacy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	14	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	2	knowledge	knowledge	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	5	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	6	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	9	understanding	understanding	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	11	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	12	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	13	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	14	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	15	handled	handle	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	16	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	17	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	18	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	19	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	20	help	help	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	21	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	22	improve	improve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	24	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	26	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	27	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	28	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	29	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	30	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	31	consequently	consequently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	32	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	33	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	15	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	2	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	3	analyzes	analyze	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	4	fifty	fifty	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	7	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	8	relevant	relevant	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	9	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	10	proposed	propose	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	11	along	along	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	12	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	13	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	15	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	16	research	research	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	17	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	18	proposes	propose	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	19	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	20	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	21	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	23	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	24	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	25	classified	classify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	26	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	27	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	29	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	30	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	31	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	32	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	16	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	1	For	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	2	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	3	group	group	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	6	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	7	advantages	advantage	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	9	shortcomings	shortcoming	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	10	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	11	highlighted	highlight	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	17	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	1	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	2	Introduction	Introduction	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	3	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	5	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	6	pursue	pursue	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	7	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	8	discover	discover	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	9	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	11	permit	permit	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	12	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	14	foresee	foresee	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	15	subsequent	subsequent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	17	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	18	demands	demand	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	18	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	1	Applications	application	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	3	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	4	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	5	include	include	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	6	recommendation	recommendation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	7	systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	8	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	9	11	11	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	10	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	12	e	e	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	13	-	-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	14	commerce	commerce	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	15	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	16	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	17	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	19	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	20	personalization	personalization	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	21	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	22	34	34	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	23	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	25	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	26	reduction	reduction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	28	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	29	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	30	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	31	latencies	latency	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	32	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	33	means	mean	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	34	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	35	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	36	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	37	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	38	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	39	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	40	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	41	cache	cache	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	42	prevalidation	prevalidation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	43	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	44	17	17	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	45	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	19	46	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	1	Since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	3	goals	goal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	5	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	6	applications	application	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	7	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	8	completely	completely	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	9	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	11	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	12	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	13	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	14	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	15	requirements	requirement	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	16	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	17	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	18	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	19	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	21	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	20	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	1	For	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	2	instance	instance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	5	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	8	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	9	might	may	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	10	work	work	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	11	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	12	object-	object-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	13	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	14	granularity	granularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	15	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	16	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	17	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	18	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	19	might	may	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	20	reduce	reduce	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	22	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	23	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	24	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	25	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	21	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	3	recommendation	recommendation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	4	systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	5	focus	focus	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	7	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	9	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	10	granularity	granularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	11	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	12	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	13	expect	expect	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	14	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	16	system	system	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	17	recommends	recommend	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	18	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	19	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	21	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	22	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	23	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	24	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	25	composing	compose	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	26	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	27	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	22	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	2	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	3	focuses	focus	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	4	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	7	applied	apply	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	8	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	10	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	23	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	2	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	5	technique	technique	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	6	aimed	aim	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	7	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	8	reducing	reduce	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	9	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	11	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	12	latencies	latency	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	13	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	14	process-	process-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	15	ing	ing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	16	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	17	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	18	downloading	download	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	19	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	20	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	21	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	22	before	before	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	23	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	24	actually	actually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	25	demand	demand	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	26	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	24	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	1	For	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	2	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	3	example	example	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	5	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	7	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	8	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	9	likely	likely	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	10	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	11	visit	visit	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	12	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	13	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	14	after	after	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	15	visiting	visit	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	16	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	17	A	A	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	20	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	21	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	22	download	download	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	24	cache	cache	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	25	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	26	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	27	just	just	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	28	after	after	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	29	requesting	request	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	30	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	31	A	a	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	32	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	33	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	34	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	35	while	while	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	36	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	37	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	38	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	39	reading	read	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	40	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	25	41	A.	a.	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	3	case	case	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	5	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	7	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	8	actually	actually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	9	requests	request	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	10	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	11	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	13	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	14	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	15	displayed	display	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	16	without	without	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	17	network	network	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	18	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	19	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	20	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	21	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	22	already	already	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	23	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	24	cache	cache	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	26	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	1	Accurate	accurate	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	2	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	4	required	require	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	5	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	8	misprediction	misprediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	10	prefetched	prefetched	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	11	objects	object	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	12	waste	waste	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	13	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	15	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	16	and/or	and/or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	17	network	network	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	18	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	20	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	21	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	22	could	could	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	23	degrade	degrade	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	25	overall	overall	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	26	sys-	sys-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	27	tem	tem	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	28	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	27	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	1	For	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	3	reason	reason	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	7	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	8	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	10	core	core	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	12	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	13	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	28	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	2	wide	wide	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	3	range	range	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	7	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	9	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	10	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	11	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	12	pro-	pro-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	13	posed	pose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	14	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	16	literature	literature	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	17	over	over	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	19	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	20	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	21	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	29	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	1	Although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	2	early	early	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	3	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	4	focused	focus	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	5	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	6	exploiting	exploit	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	8	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	9	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	10	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	11	determine	determine	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	13	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	14	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	16	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	17	likely	likely	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	18	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	19	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	20	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	21	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	23	near	near	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	24	future	future	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	25	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	27	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	28	explored	explored	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	29	option	option	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	30	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	31	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	32	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	33	subsequent	subsequent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	34	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	35	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	36	past	past	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	37	sequences	sequence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	38	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	39	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	30	40	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	3	over	over	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	5	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	6	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	8	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	9	data	data	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	10	sources	source	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	11	beyond	beyond	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	13	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	15	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	17	such	such	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	18	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	20	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	21	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	22	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	25	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	26	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	27	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	29	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	30	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	31	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	32	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	33	being	be	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	34	included	include	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	35	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	36	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	37	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	38	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	31	39	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	1	Understanding	understanding	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	3	identifying	identify	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	4	advantages	advantage	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	6	shortcomings	shortcoming	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	9	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	10	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	11	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	12	hard	hard	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	13	task	task	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	14	even	even	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	15	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	16	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	17	experienced	experienced	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	18	researcher	researcher	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	19	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	20	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	21	field	field	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	22	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	23	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	25	large	large	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	26	amount	amount	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	28	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	29	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	30	work	work	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	31	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	32	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	33	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	34	rules	rule	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	32	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	2	means	mean	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	3	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	4	selecting	select	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	6	best	good	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	7	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	8	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	9	developing	develop	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	10	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	11	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	12	one	one	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	13	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	14	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	15	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	16	environment	environment	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	17	represents	represent	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	18	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	19	arduous	arduous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	20	task	task	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	33	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	1	Few	few	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	2	attempts	attempt	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	3	classifying	classify	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	5	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	6	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	7	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	8	published	publish	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	10	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	11	mainly	mainly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	12	focus	focus	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	13	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	14	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	15	sequences	sequence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	17	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	18	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	19	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	34	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	1	Rabinovich	Rabinovich	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	3	Spatscheck	Spatscheck	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	4	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	5	41	41	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	6	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	7	classify	classify	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	8	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	9	approaches	approach	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	10	depending	depend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	11	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	12	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	13	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	14	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	15	made	make	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	35	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	1	They	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	2	differentiate	differentiate	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	3	three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	4	categories	category	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	6	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	7	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	8	depending	depend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	10	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	11	many	many	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	12	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	13	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	14	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	15	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	36	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	1	Other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	2	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	3	includes	include	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	4	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	8	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	9	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	10	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	11	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	12	exhibit	exhibit	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	15	another	another	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	16	one	one	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	17	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	18	dedicated	dedicate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	19	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	20	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	21	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	22	exploiting	exploit	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	24	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	27	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	37	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	2	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	3	classification	classification	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	4	published	publish	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	6	presented	present	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	7	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	8	Davison	Davison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	9	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	10	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	11	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	13	divides	divide	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	15	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	16	algorithms	algorithm	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	17	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	18	five	five	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	19	categories	category	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	20	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	22	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	24	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	26	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	27	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	28	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	38	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	1	Although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	2	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	4	mentioned	mention	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	5	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	6	more	more	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	7	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	8	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	9	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	10	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	11	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	12	extra	extra	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	13	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	14	sources	source	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	16	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	17	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	18	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	20	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	21	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	22	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	23	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	25	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	26	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	27	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	28	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	29	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	30	depth	depth	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	39	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	1	Unfortunately	unfortunately	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	4	complexity	complexity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	7	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	8	published	publish	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	9	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	11	liter-	liter-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	12	ature	ature	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	13	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	14	Davison	Davison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	15	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	16	classification	classification	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	17	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	18	noticeably	noticeably	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	19	increased	increase	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	40	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	2	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	3	consequence	consequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	5	many	many	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	7	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	8	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	9	do	do	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	10	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	11	fit	fit	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	12	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	13	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	14	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	17	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	18	identified	identify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	19	cate-	cate-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	20	gories	gorie	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	41	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	1	Therefore	therefore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	3	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	4	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	5	effort	effort	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	7	required	require	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	8	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	9	order	order	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	11	clarify	clarify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	12	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	14	provide	provide	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	15	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	16	sound	sound	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	17	understanding	understanding	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	19	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	20	advantages	advantage	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	22	progresses	progress	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	24	real	real	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	25	possibilities	possibility	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	27	use	use	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	42	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	3	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	5	we	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	6	propose	propose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	7	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	9	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	10	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	11	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	12	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	14	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	16	classified	classify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	17	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	19	type	type	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	21	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	22	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	23	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	24	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	25	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	26	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	43	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	4	organized	organize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	6	four	four	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	7	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	8	categories	category	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	9	depending	depend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	11	whether	whether	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	12	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	13	take	take	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	14	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	15	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	17	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	18	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	21	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	22	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	23	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	26	HTTP	HTTP	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	27	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	28	and/or	and/or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	29	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	30	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	31	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	32	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	33	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	44	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	2	classification	classification	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	4	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	5	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	7	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	10	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	11	determine	determine	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	12	what	what	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	13	kind	kind	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	15	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	16	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	17	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	18	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	19	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	20	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	21	complex	complex	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	23	analysis	analysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	24	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	45	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	1	For	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	2	instance	instance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	5	recently	recently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	6	added	add	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	7	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	8	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	9	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	10	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	11	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	12	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	13	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	14	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	15	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	16	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	18	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	20	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	21	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	22	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	23	at	at	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	24	least	least	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	25	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	26	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	27	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	28	already	already	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	29	followed	follow	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	30	that	that	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	31	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	46	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	2	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	3	contribution	contribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	5	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	6	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	7	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	8	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	9	present	present	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	10	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	11	comprehensive	comprehensive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	12	review	review	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	14	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	15	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	16	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	17	grouped	group	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	18	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	19	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	20	novel	novel	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	21	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	23	thus	thus	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	24	easing	ease	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	26	understanding	understanding	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	28	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	29	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	30	work	work	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	31	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	33	implications	implication	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	34	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	35	selecting	select	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	36	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	37	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	38	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	39	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	40	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	41	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	42	designing	design	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	43	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	44	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	45	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	46	system	system	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	47	47	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	2	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	3	words	word	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	5	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	6	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	7	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	8	aimed	aim	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	9	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	10	examining	examine	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	12	pros	pro	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	14	cons	con	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	16	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	17	group	group	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	19	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	20	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	21	identifying	identify	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	22	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	24	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	25	trends	trend	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	26	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	27	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	28	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	29	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	48	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	2	remainder	remainder	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	5	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	7	organized	organize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	8	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	9	follows	follow	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	49	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	1	Section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	2	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	3	describes	describe	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	5	elements	element	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	6	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	7	compose	compose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	10	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	11	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	13	highlighting	highlight	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	15	func-	func-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	16	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	19	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	20	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	50	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	1	Section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	2	3	3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	3	classifies	classifie	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	4	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	7	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	8	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	10	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	11	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	51	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	1	Finally	finally	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	3	Section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	4	4	4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	5	draws	draw	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	6	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	7	concluding	conclude	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	8	remarks	remark	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	52	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	1	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	2	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	3	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	4	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	5	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	6	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	7	systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	8	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	9	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	10	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	11	extending	extend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	13	generic	generic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	14	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	15	archi-	archi-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	16	tecture	tecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	17	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	18	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	19	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	21	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	22	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	23	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	24	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	25	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	26	two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	27	extra	extra	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	28	elements	element	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	29	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	30	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	31	predic-	predic-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	32	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	33	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	34	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	35	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	36	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	53	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	2	function	function	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	6	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	7	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	8	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	9	guess	guess	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	10	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	11	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	12	actions	action	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	14	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	16	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	54	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	1	Its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	2	implementation	implementation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	3	corresponds	correspond	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	4	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	7	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	8	whose	whose	WP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	9	output	output	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	10	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	11	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	12	hint	hint	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	13	list	list	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	55	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	2	most	most	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	3	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	5	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	6	hint	hint	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	7	list	list	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	8	includes	include	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	9	references	reference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	10	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	12	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	13	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	15	although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	16	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	17	few	few	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	18	approaches	approach	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	19	suggest	suggest	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	20	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	21	type	type	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	22	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	23	actions	action	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	24	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	25	12	12	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	26	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	56	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	2	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	3	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	4	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	5	hints	hint	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	6	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	7	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	8	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	9	made	make	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	10	basing	base	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	11	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	12	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	13	experience	experience	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	14	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	15	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	16	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	17	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	18	preferences	preference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	57	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	2	hints	hint	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	4	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	5	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	7	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	8	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	10	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	11	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	12	aimed	aim	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	13	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	14	preprocessing	preprocesse	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	15	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	16	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	18	downloading	download	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	19	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	20	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	21	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	22	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	23	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	25	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	26	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	58	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	1	By	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	2	preprocess-	preprocess-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	3	ing	e	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	5	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	7	advance	advance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	10	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	11	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	12	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	13	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	14	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	15	reduced	reduce	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	16	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	17	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	18	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	19	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	20	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	21	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	22	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	59	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	2	amount	amount	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	3	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	4	scope	scope	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	7	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	8	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	9	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	10	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	11	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	12	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	14	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	15	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	16	depends	depend	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	17	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	19	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	22	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	23	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	24	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	26	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	27	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	29	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	30	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	31	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	32	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	33	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	34	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	35	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	36	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	37	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	38	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	39	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	40	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	60	41	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	1	When	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	3	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	4	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	6	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	7	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	9	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	12	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	13	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	14	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	15	use	use	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	16	access	access	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	17	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	19	con-	con-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	20	tents	tent	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	21	seen	see	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	22	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	23	preferences	preference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	24	coming	come	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	25	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	26	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	27	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	28	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	61	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	1	Consequently	consequently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	3	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	4	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	5	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	6	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	7	useful	useful	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	8	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	9	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	10	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	62	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	1	If	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	3	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	4	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	6	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	7	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	9	proxy	proxy	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	10	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	12	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	13	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	14	take	take	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	15	advantage	advantage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	18	multi	multi	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	19	-	-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	20	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	22	multi	multi	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	23	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	24	server	server	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	25	informa-	informa-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	26	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	27	gathered	gather	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	28	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	29	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	30	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	31	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	32	perform	perform	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	33	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	34	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	63	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	1	When	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	3	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	5	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	6	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	8	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	9	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	10	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	11	origin	origin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	12	server	server	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	13	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	14	replica	replica	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	17	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	19	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	20	Dis-	Dis-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	21	tribution	tribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	22	Networks	Networks	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	23	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	24	CDN	CDN	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	25	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	26	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	27	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	28	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	29	makes	make	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	30	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	31	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	32	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	33	multi	multi	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	34	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	35	user	user	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	36	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	37	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	38	preferences	preference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	39	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	40	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	41	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	42	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	43	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	64	44	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	2	option	option	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	3	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	4	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	6	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	7	explored	explore	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	10	research	research	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	11	literature	literature	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	12	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	15	accuracy	accuracy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	18	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	19	made	make	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	20	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	21	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	22	potential	potential	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	23	use	use	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	24	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	25	real	real	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	26	scenarios	scenario	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	65	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	1	Finally	finally	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	3	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	4	more	more	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	5	complex	complex	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	6	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	9	pre-	pre-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	10	dictions	diction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	11	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	12	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	13	performed	perform	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	14	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	15	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	16	elements	element	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	17	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	18	collaboration	collaboration	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	66	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	2	benefits	benefit	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	4	accuracy	accuracy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	6	coverage	coverage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	9	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	67	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	2	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	3	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	5	independent	independent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	6	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	8	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	10	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	11	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	12	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	13	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	14	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	15	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	18	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	19	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	20	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	21	receives	receive	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	23	hint	hint	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	24	list	list	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	68	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	4	common	common	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	5	trend	trend	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	7	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	8	locate	locate	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	9	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	10	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	12	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	13	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	14	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	15	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	16	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	17	currently	currently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	18	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	19	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	20	Mozilla	Mozilla	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	21	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	22	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	23	browsers	browser	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	69	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	2	prefetched	prefetched	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	3	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	4	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	5	stored	store	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	7	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	8	cache	cache	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	9	until	until	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	10	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	11	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	12	demanded	demand	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	13	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	14	evicted	evict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	70	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	1	Therefore	therefore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	3	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	4	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	5	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	6	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	7	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	8	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	9	stored	store	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	10	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	12	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	13	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	14	must	must	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	15	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	16	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	17	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	18	prefetched	prefetche	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	71	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	3	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	5	drawback	drawback	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	6	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	8	although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	10	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	11	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	12	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	13	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	15	result	result	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	17	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	18	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	19	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	21	many	many	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	22	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	24	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	25	composing	compose	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	26	that	that	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	27	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	28	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	29	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	30	static	static	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	31	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	32	consequently	consequently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	33	cacheable	cacheable	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	72	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	1	Furthermore	furthermore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	3	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	4	dynamically	dynamically	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	5	generated	generate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	6	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	7	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	8	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	9	cached	cache	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	10	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	11	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	12	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	13	properly	properly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	14	mark	mark	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	16	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	73	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	2	all	all	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	4	cacheable	cacheable	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	5	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	6	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	7	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	8	prefetched	prefetche	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	9	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	10	precached	precache	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	11	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	13	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	14	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	15	3	3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	16	Figure	figure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	17	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	18	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	19	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	20	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	21	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	22	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	23	cover	cover	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	25	among	among	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	26	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	27	technologies	technology	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	28	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	29	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	30	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	31	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	32	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	33	application	application	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	34	program-	program-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	35	ming	ming	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	36	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	37	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	38	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	39	application	application	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	40	programming	programming	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	41	like	like	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	42	AJAX	AJAX	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	43	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	44	Java	Java	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	45	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	46	Flash	Flash	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	47	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	48	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	49	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	50	Rich	Rich	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	51	Internet	Internet	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	52	Applications	Applications	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	74	53	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	2	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	5	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	7	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	8	measured	measure	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	9	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	10	means	mean	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	12	metrics	metric	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	13	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	14	quantify	quantify	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	15	both	both	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	17	efficiency	efficiency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	20	efficacy	efficacy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	23	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	75	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	1	Although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	2	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	3	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	4	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	5	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	6	specific	specific	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	7	indexes	index	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	10	general	general	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	11	trend	trend	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	12	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	13	evaluating	evaluate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	14	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	15	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	16	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	17	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	18	trade	trade	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	19	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	20	off	off	RP
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	21	Recall	Recall	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	22	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	23	Precision	Precision	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	25	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	26	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	27	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	28	metrics	metric	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	76	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	1	Recall	recall	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	3	defined	define	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	4	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	6	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	8	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	9	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	11	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	12	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	13	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	77	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	3	recall	recall	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	4	quantifies	quantify	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	6	weight	weight	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	9	predicted	predict	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	10	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	11	over	over	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	13	amount	amount	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	15	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	16	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	17	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	19	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	21	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	22	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	23	give	give	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	24	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	25	idea	idea	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	26	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	28	usefulness	usefulness	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	29	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	30	coverage	coverage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	31	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	33	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	78	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	1	Precision	precision	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	3	defined	define	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	4	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	6	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	8	good	good	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	9	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	10	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	12	number	number	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	14	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	16	so	so	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	17	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	18	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	19	related	relate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	20	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	22	quantity	quantity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	23	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	24	wasted	waste	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	25	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	79	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	1	3	3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	2	Prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	3	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	4	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	5	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	6	section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	7	presents	present	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	9	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	10	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	12	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	13	breaks	break	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	14	down	down	RP
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	16	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	17	algorithms	algorithm	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	18	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	19	four	four	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	20	categories	category	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	21	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	22	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	24	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	25	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	26	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	27	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	29	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	80	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	1	Earlier	early	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	2	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	3	simply	simply	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	4	consider	consider	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	5	isolated	isolated	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	6	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	7	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	9	while	while	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	10	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	11	oth-	oth-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	12	ers	er	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	13	analyze	analyze	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	15	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	17	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	18	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	19	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	20	discover	discover	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	21	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	22	navigation	navigation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	23	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	81	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	2	discover	discover	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	4	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	7	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	10	relation	relation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	11	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	12	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	14	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	15	re-	re-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	16	cent	cent	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	17	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	18	take	take	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	19	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	20	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	22	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	23	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	24	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	25	examine	examine	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	26	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	27	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	82	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	1	Content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	3	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	4	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	5	may	may	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	6	involve	involve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	8	analysis	analysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	11	list	list	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	13	hyperlinks	hyperlinks	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	16	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	17	surrounding	surround	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	18	hyperlinks	hyperlink	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	19	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	21	labels	label	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	22	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	23	meta	meta	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	24	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	25	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	26	included	include	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	27	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	29	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	30	designer	designer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	83	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	2	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	4	represented	represent	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	6	Figure	Figure	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	7	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	9	summarizes	summarize	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	11	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	12	algo-	algo-	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	13	rithms	rithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	14	ranging	range	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	15	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	17	earliest	early	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	18	works	work	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	19	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	20	1995	1995	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	23	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	24	recent	recent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	25	ones	one	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	26	found	find	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	27	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	28	2010	2010	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	84	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	3	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	4	designed	design	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	5	considering	consider	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	7	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	9	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	10	contemporary	contemporary	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	11	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	12	sites	site	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	14	internet	internet	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	15	technologies	technology	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	85	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	3	sense	sense	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	5	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	6	survey	survey	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	7	represents	represent	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	9	review	review	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	11	covers	cover	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	12	more	more	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	13	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	14	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	15	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	17	research	research	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	18	following	follow	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	20	evolution	evolution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	23	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	25	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	27	genesis	genesis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	28	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	29	static	static	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	30	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	31	had	have	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	32	few	few	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	33	embedded	embed	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	34	ob-	ob-	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	35	jects	ject	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	36	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	37	hyperlinks	hyperlinks	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	38	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	39	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	40	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	41	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	42	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	43	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	44	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	45	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	46	personalized	personalized	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	47	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	86	48	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	1	4	4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	2	3.1	3.1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	3	Object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	4	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	5	Most	Most	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	8	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	9	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	10	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	11	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	12	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	13	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	14	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	15	early	early	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	18	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	87	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	2	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	3	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	5	more	more	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	6	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	7	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	8	decade	decade	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	9	ago	ago	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	12	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	13	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	14	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	15	much	much	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	16	simpler	simple	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	17	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	18	nowadays	nowadays	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	88	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	2	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	3	consequence	consequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	8	absence	absence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	10	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	11	approaches	approach	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	14	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	15	intuitive	intuitive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	17	simplest	simplest	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	18	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	19	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	20	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	89	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	2	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	3	focus	focus	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	4	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	5	analyzing	analyze	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	6	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	7	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	9	like	like	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	10	temporal	temporal	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	11	lo-	lo-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	12	cality	cality	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	14	frequency	frequency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	16	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	17	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	18	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	20	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	21	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	23	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	24	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	26	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	27	aspects	aspect	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	28	studied	study	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	29	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	30	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	31	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	90	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	1	Algorithms	Algorithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	2	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	3	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	5	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	6	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	7	follow	follow	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	9	long-	long-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	10	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	11	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	13	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	14	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	16	subscribed	subscribe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	17	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	18	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	19	storing	store	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	20	specific	specific	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	21	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	23	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	24	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	25	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	26	charge	charge	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	28	keeping	keep	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	29	updated	update	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	30	copies	copy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	31	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	33	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	91	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	2	opposite	opposite	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	4	referred	refer	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	5	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	6	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	8	short	short	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	9	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	10	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	11	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	13	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	14	aims	aim	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	15	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	16	prefetch	prefetch	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	17	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	18	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	19	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	20	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	21	likely	likely	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	22	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	23	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	24	referenced	reference	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	25	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	27	near	near	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	28	future	future	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	29	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	30	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	31	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	32	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	33	recent	recent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	34	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	92	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	2	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	3	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	4	Bestavros	Bestavros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	5	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	6	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	93	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	2	5	5	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	4	provides	provide	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	5	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	6	overview	overview	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	9	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	10	benefits	benefit	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	12	propagating	propagate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	13	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	14	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	15	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	16	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	17	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	18	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	19	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	20	closer	close	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	22	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	23	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	94	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	1	No	no	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	4	described	describe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	5	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	6	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	7	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	8	ideas	idea	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	10	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	12	al-	al-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	13	gorithm	gorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	14	should	should	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	15	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	16	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	95	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	2	suggested	suggested	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	4	takes	take	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	5	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	6	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	8	rate	rate	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	10	transferred	transfer	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	11	bytes	byte	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	12	among	among	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	13	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	15	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	18	cumulative	cumulative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	19	distribution	distribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	21	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	22	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	24	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	25	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	26	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	96	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	2	idea	idea	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	3	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	4	deeply	deeply	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	5	worked	work	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	7	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	8	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	9	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	11	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	12	author	author	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	13	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	14	6	6	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	15	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	17	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	18	explores	explore	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	20	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	21	gains	gain	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	22	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	23	propagating	propagate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	24	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	25	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	26	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	27	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	28	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	97	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	1	That	that	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	2	is	is	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	5	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	7	replicated	replicate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	8	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	9	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	10	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	11	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	12	closer	close	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	13	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	15	final	final	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	16	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	98	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	2	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	4	propagation	propagation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	6	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	9	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	12	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	14	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	16	expected	expect	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	17	reduction	reduction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	18	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	19	traffic	traffic	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	99	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	2	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	3	exploits	exploit	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	5	properties	property	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	7	temporal	temporal	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	9	geographical	geographical	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	10	locality	locality	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	12	ref-	ref-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	13	erences	erence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	15	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	17	unlike	unlike	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	19	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	20	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	22	require	require	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	23	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	24	large	large	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	25	community	community	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	26	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	27	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	28	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	29	thousands	thousand	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	30	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	31	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	32	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	100	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	2	Top-10	Top-10	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	3	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	4	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	5	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	6	Markatos	Markatos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	8	Chronaki	Chronaki	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	9	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	10	33	33	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	11	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	13	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	14	pub-	pub-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	15	lished	lishe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	18	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	19	century	century	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	21	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	22	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	23	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	25	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	26	attempts	attempt	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	27	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	28	exploit	exploit	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	29	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	30	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	31	doc-	doc-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	32	uments	ument	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	33	’	’	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	34	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	35	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	36	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	37	purposes	purpose	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	101	38	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	1	Additional	additional	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	2	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	3	obtained	obtain	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	4	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	6	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	7	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	8	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	9	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	10	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	11	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	12	adapt	adapt	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	13	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	14	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	15	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	16	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	17	profiles	profile	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	102	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	3	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	7	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	8	publishes	publish	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	10	list	list	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	13	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	14	popular	popular	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	15	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	17	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	18	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	19	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	20	complemented	complement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	21	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	22	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	23	access	access	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	24	preferences	preference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	25	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	26	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	27	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	28	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	29	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	103	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	2	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	3	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	4	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	5	included	include	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	7	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	8	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	9	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	11	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	12	idea	idea	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	15	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	16	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	18	use	use	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	21	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	22	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	23	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	24	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	25	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	26	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	27	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	28	although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	29	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	30	profile	profile	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	31	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	32	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	33	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	34	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	35	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	36	control	control	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	37	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	38	adapt	adapt	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	39	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	104	40	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	3	26	26	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	6	Jian	Jian	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	7	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	8	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	105	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	1	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	2	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	4	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	5	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	6	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	7	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	8	exhibit-	exhibit-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	9	ing	e	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	10	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	11	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	12	characteristic	characteristic	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	106	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	3	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	5	three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	6	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	7	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	8	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	9	explored	explore	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	11	quantified	quantify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	12	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	13	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	14	lifetime	lifetime	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	16	probability	probability	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	18	being	be	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	19	accessed	access	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	22	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	23	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	24	arrival	arrival	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	25	rate	rate	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	107	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	3	prefetches	prefetche	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	4	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	5	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	6	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	8	estimated	estimate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	9	val-	val-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	10	ues	ues	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	11	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	12	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	13	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	14	characteristic	characteristic	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	15	exceed	exceed	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	16	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	17	specific	specific	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	18	threshold	threshold	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	108	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	2	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	3	works	work	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	4	among	among	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	5	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	8	either	either	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	9	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	10	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	11	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	109	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	2	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	3	collects	collect	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	4	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	5	characterizing	characterize	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	7	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	9	makes	make	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	10	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	11	available	available	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	12	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	14	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	15	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	16	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	17	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	110	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	111	1	Wu	Wu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	111	2	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	111	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	111	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	2	47	47	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	4	propose	propose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	5	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	6	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	7	aimed	aim	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	8	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	9	improving	improve	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	10	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	11	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	12	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	13	metric	metric	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	15	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	17	hit	hit	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	18	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	20	bandwidth	bandwidth	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	21	consumption	consumption	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	23	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	24	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	25	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	27	both	both	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	28	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	29	H	h	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	30	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	31	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	32	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	112	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	1	De-	de-	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	2	pending	pende	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	3	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	5	target	target	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	6	metric	metric	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	9	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	10	prefetches	prefetche	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	11	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	12	showing	show	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	13	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	14	particular	particular	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	15	characteristic	characteristic	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	16	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	17	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	19	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	21	longer	long	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	22	update	update	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	23	intervals	interval	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	24	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	25	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	26	mix	mix	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	28	both	both	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	29	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	113	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	2	best	good	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	3	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	5	achieved	achieve	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	6	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	8	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	9	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	10	maximizes	maximize	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	12	H	H	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	13	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	14	B	B	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	15	metric	metric	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	114	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	2	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	3	assumes	assume	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	4	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	6	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	7	push	push	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	8	fresh	fresh	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	9	copies	copy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	10	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	11	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	12	5	5	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	13	caches	cache	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	14	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	15	assuming	assume	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	16	unlimited	unlimited	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	17	cache	cache	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	18	size	size	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	19	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	20	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	21	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	22	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	23	whenever	whenever	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	24	such	such	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	25	copies	copy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	26	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	27	updated	update	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	115	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	3	46	46	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	6	Venkataramani	Venkataramani	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	7	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	8	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	116	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	1	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	2	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	3	long	long	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	4	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	5	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	6	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	7	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	8	proxy	proxy	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	9	caches	cache	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	10	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	11	subscribed	subscribe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	12	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	13	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	14	popular	popular	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	15	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	16	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	17	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	18	updated	update	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	19	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	20	low	low	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	21	fre-	fre-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	22	quency	quency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	117	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	2	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	4	aimed	aim	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	5	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	6	being	be	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	7	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	10	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	11	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	12	caches	cache	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	14	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	15	CDN	CDN	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	16	hierarchical	hierarchical	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	17	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	18	while	while	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	20	typical	typical	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	21	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	22	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	23	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	25	short	short	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	26	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	27	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	28	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	29	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	30	should	should	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	31	remain	remain	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	32	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	33	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	34	lower	low	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	35	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	36	caches	cache	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	118	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	3	30	30	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	6	Lau	Lau	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	8	Ng	ng	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	9	propose	propose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	10	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	11	prefet-	prefet-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	12	ching	ching	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	13	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	14	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	15	analyzes	analyze	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	17	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	18	’s	’s	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	19	cache	cache	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	20	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	23	long	long	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	24	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	25	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	26	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	119	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	3	predicts	predict	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	5	prefetches	prefetche	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	6	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	7	documents	document	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	9	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	10	linked	link	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	11	documents	document	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	12	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	13	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	120	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	1	Two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	2	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	3	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	4	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	5	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	7	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	8	ob-	ob-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	9	ject	ject	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	10	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	11	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	12	access	access	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	13	rate	rate	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	16	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	17	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	18	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	19	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	20	accessed	access	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	121	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	2	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	3	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	4	26	26	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	5	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	7	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	8	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	9	whose	whose	WP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	10	characteristic	characteristic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	11	exceeds	exceed	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	12	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	13	given	give	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	14	threshold	threshold	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	16	prefetched	prefetche	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	122	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	3	uses	use	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	5	detection	detection	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	6	theory	theory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	7	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	8	determine	determine	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	10	threshold	threshold	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	11	value	value	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	123	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	2	addition	addition	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	4	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	5	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	6	take	take	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	7	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	8	active	active	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	9	part	part	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	10	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	12	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	13	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	14	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	15	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	16	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	17	graphic	graphic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	18	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	19	interface	interface	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	20	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	21	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	22	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	23	modify	modify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	24	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	25	computed	compute	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	26	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	27	schedule	schedule	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	124	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	2	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	3	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	4	Rangarajan	Rangarajan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	5	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	6	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	125	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	2	42	42	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	4	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	5	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	6	characterized	characterize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	7	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	9	short	short	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	11	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	12	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	126	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	2	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	3	presents	present	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	5	technique	technique	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	6	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	7	dynamically	dynamically	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	8	groups	group	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	9	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	10	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	11	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	12	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	13	frequency	frequency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	15	access	access	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	16	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	17	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	18	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	127	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	3	way	way	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	6	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	7	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	8	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	9	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	10	reorganized	reorganize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	11	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	12	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	14	needs	need	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	16	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	17	group	group	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	19	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	128	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	2	avoid	avoid	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	3	overloading	overload	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	5	network	network	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	8	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	9	predicts	predict	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	10	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	11	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	12	cluster	cluster	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	13	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	14	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	15	fashion	fashion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	16	instead	instead	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	18	predicting	predict	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	19	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	20	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	21	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	22	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	23	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	24	that	that	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	25	is	is	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	26	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	27	whenever	whenever	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	28	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	29	host	host	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	30	connects	connect	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	31	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	33	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	34	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	35	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	36	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	37	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	38	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	39	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	40	strategy	strategy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	41	returns	return	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	42	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	43	URIs	URIs	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	44	predicted	predict	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	45	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	46	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	47	cluster	cluster	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	48	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	49	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	50	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	51	host	host	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	52	belongs	belong	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	129	53	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	1	Unlike	unlike	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	3	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	4	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	8	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	9	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	10	Duchamp	Duchamp	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	11	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	12	21	21	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	13	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	15	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	16	manage	manage	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	17	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	18	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	19	statistics	statistic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	20	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	22	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	23	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	24	store	store	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	25	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	26	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	27	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	28	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	29	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	30	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	31	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	130	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	1	All	all	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	3	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	5	then	then	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	6	summarized	summarize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	9	probability	probability	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	11	being	be	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	12	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	13	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	14	estimated	estimate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	131	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	2	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	4	takes	take	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	5	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	6	both	both	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	7	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	9	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	11	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	12	sides	side	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	132	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	1	Once	once	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	3	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	5	received	receive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	8	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	9	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	10	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	11	interact	interact	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	12	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	13	decide	decide	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	14	whether	whether	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	15	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	16	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	17	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	18	prefetch	prefetch	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	20	hinted	hinted	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	21	hyperlinks	hyperlinks	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	22	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	23	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	24	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	25	local	local	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	26	parameters	parameter	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	133	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	1	Algorithms	Algorithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	2	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	4	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	5	made	make	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	6	sense	sense	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	7	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	8	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	9	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	10	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	11	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	12	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	14	early	early	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	17	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	18	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	20	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	22	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	23	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	24	simple	simple	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	25	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	26	static	static	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	27	regarding	regard	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	29	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	30	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	31	point	point	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	32	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	33	view	view	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	34	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	35	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	36	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	37	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	38	lifetime	lifetime	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	134	39	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	1	Nowadays	nowadays	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	4	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	5	fits	fit	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	6	better	well	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	7	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	8	CDNs	CDNs	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	10	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	11	control	control	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	12	both	both	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	14	origin	origin	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	15	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	16	replica	replica	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	17	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	135	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	1	3.2	3.2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	2	Sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	4	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	5	Most	Most	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	9	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	10	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	11	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	14	literature	literature	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	15	lie	lie	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	18	anal-	anal-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	19	ysis	ysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	22	paths	path	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	23	followed	follow	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	24	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	25	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	26	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	27	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	28	that	that	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	29	is	is	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	30	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	31	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	33	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	34	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	35	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	136	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	1	By	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	2	comparing	compare	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	4	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	5	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	6	recent	recent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	7	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	8	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	10	paths	path	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	11	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	12	recorded	record	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	14	algo-	algo-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	15	rithms	rithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	16	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	17	able	able	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	18	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	19	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	20	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	21	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	137	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	1	Algorithms	Algorithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	2	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	3	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	5	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	6	require	require	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	7	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	8	distinguish	distinguish	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	9	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	10	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	11	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	12	issuing	issue	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	14	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	16	order	order	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	17	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	18	keep	keep	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	19	track	track	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	22	right	right	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	23	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	25	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	138	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	1	Al-	al-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	2	though	though	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	3	solving	solve	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	5	issue	issue	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	7	trivial	trivial	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	8	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	10	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	11	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	12	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	13	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	14	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	17	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	19	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	20	becomes	become	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	21	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	22	complex	complex	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	23	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	24	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	25	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	26	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	27	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	29	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	30	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	31	keeping	keep	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	32	track	track	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	33	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	34	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	35	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	36	IP	IP	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	37	could	could	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	38	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	39	enough	enough	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	40	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	41	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	42	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	43	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	44	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	45	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	46	frequently	frequently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	47	requiring	require	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	48	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	49	use	use	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	50	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	51	cookies	cookie	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	52	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	139	53	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	2	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	6	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	7	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	8	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	10	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	12	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	13	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	14	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	15	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	17	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	19	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	20	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	21	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	22	implemented	implement	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	23	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	24	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	25	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	26	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	27	limiting	limit	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	28	factor	factor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	29	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	30	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	31	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	32	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	33	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	34	type	type	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	35	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	36	19	19	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	37	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	38	29	29	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	39	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	140	40	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	3	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	4	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	5	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	7	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	8	6	6	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	9	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	10	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	11	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	12	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	13	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	14	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	15	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	16	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	17	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	18	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	19	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	20	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	21	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	22	before	before	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	23	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	25	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	26	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	28	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	29	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	30	placed	place	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	31	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	32	first	first	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	33	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	34	seen	see	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	35	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	36	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	141	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	2	expected	expect	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	5	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	7	first	first	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	9	seen	see	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	10	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	11	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	12	low	low	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	13	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	14	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	16	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	17	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	19	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	20	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	21	high	high	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	22	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	23	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	24	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	26	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	27	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	28	19	19	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	29	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	142	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	4	limitation	limitation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	5	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	6	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	7	overcome	overcome	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	8	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	9	additional	additional	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	10	sources	source	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	12	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	13	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	14	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	15	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	17	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	18	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	143	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	1	Therefore	therefore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	3	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	5	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	6	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	7	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	8	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	9	divided	divide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	10	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	11	pure	pure	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	13	hybrid	hybrid	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	15	depending	depend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	16	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	17	whether	whether	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	19	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	20	considers	consider	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	21	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	23	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	25	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	26	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	27	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	28	takes	take	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	29	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	30	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	31	additional	additional	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	32	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	33	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	34	refine	refine	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	35	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	144	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	1	3.2.1	3.2.1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	2	Pure	pure	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	3	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	4	There	there	EX
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	7	large	large	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	8	number	number	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	10	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	11	specifically	specifically	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	12	addressed	address	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	14	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	15	subse-	subse-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	16	quent	quent	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	17	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	18	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	20	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	22	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	23	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	24	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	25	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	26	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	27	receives	receive	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	145	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	1	Some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	2	variations	variation	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	4	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	5	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	6	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	7	often	often	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	8	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	10	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	12	hint	hint	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	13	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	14	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	15	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	16	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	19	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	20	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	21	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	22	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	25	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	26	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	28	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	29	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	30	will	will	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	31	prefetch	prefetch	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	32	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	33	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	34	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	35	13	13	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	36	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	37	4	4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	38	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	39	38	38	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	40	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	146	41	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	2	dependency	dependency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	3	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	4	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	5	DG	DG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	6	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	7	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	8	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	9	Padmanabhan	Padmanabhan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	11	Mogul	Mogul	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	12	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	13	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	14	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	15	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	16	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	17	taken	take	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	18	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	19	baseline	baseline	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	20	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	21	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	22	comparisons	comparison	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	147	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	2	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	3	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	5	node	node	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	7	every	every	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	8	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	9	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	10	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	11	ever	ever	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	12	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	13	accessed	access	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	148	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	1	There	there	EX
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	3	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	4	arc	arc	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	5	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	6	node	node	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	7	A	a	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	8	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	9	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	10	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	12	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	13	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	14	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	15	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	16	point	point	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	17	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	18	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	19	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	20	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	21	accessed	access	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	22	B	B	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	23	within	within	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	24	w	w	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	25	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	26	after	after	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	27	A	A	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	28	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	29	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	30	w	w	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	31	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	33	lookahead	lookahead	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	34	window	window	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	35	size	size	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	149	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	2	weight	weight	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	5	arc	arc	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	8	ratio	ratio	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	11	number	number	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	13	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	14	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	15	B	b	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	16	within	within	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	17	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	18	window	window	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	19	after	after	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	20	A	a	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	23	number	number	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	25	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	26	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	27	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	28	itself	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	150	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	1	Some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	2	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	3	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	4	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	5	43	43	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	7	31	31	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	8	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	9	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	11	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	12	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	13	preprocess	preprocess	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	14	resources	resource	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	15	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	17	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	18	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	151	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	152	1	Schechter	Schechter	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	152	2	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	152	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	152	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	2	43	43	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	4	use	use	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	5	session	session	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	6	paths	path	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	7	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	8	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	9	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	10	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	11	generating	generate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	12	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	13	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	14	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	16	advance	advance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	153	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	3	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	6	distributed	distribute	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	8	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	9	cluster	cluster	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	12	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	14	Lee	Lee	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	15	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	16	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	154	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	2	31	31	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	4	combines	combine	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	7	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	8	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	10	workload	workload	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	11	distribution	distribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	12	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	14	prefetch	prefetch	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	15	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	16	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	17	disk	disk	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	18	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	19	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	20	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	21	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	155	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	156	1	Kim	Kim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	156	2	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	156	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	156	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	2	28	28	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	4	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	6	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	7	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	8	takes	take	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	9	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	10	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	12	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	14	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	15	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	17	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	18	perspective	perspective	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	157	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	2	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	4	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	6	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	8	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	9	likely	likely	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	10	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	11	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	13	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	14	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	15	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	17	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	18	will	will	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	19	follow	follow	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	158	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	1	Analyzing	analyze	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	3	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	5	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	6	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	8	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	9	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	10	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	11	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	12	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	13	advantage	advantage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	14	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	15	no	no	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	16	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	17	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	18	hidden	hide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	19	due	due	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	20	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	22	effects	effect	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	23	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	24	caching	caching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	159	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	3	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	4	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	5	considers	consider	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	6	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	7	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	8	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	10	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	11	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	12	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	13	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	14	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	15	followed	follow	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	16	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	18	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	19	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	20	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	21	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	22	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	23	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	160	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	1	Focusing	focus	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	2	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	4	proxy	proxy	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	5	perspective	perspective	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	7	Fan	Fan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	8	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	9	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	161	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	2	22	22	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	4	employ	employ	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	6	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	7	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	8	known	know	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	9	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	10	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	11	partial	partial	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	12	matching	matching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	13	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	14	PPM	PPM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	15	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	16	algorithm	algorithm	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	17	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	18	such	such	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	19	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	20	way	way	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	21	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	22	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	23	aggregates	aggregate	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	24	all	all	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	25	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	26	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	27	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	28	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	29	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	30	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	31	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	32	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	33	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	34	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	35	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	36	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	37	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	38	proxy	proxy	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	39	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	162	40	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	2	joint	joint	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	3	benefits	benefit	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	5	caching	cache	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	7	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	8	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	9	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	10	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	11	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	12	Teng	Teng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	13	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	14	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	163	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	2	45	45	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	6	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	7	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	8	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	9	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	164	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	1	Its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	2	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	7	rules	rule	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	9	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	10	one	one	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	11	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	12	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	13	confidence	confidence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	14	level	level	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	17	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	19	next	next	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	20	requested	requested	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	21	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	22	depends	depend	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	23	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	25	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	27	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	28	recently	recently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	29	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	165	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	2	contrast	contrast	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	6	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	8	Huang	Huang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	9	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	10	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	166	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	2	25	25	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	7	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	8	tries	try	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	10	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	11	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	12	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	13	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	15	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	16	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	17	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	18	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	19	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	167	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	1	Only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	2	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	3	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	5	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	6	possible	possible	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	9	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	10	uses	use	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	12	aggregate	aggregate	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	13	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	15	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	16	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	168	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	1	Since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	3	range	range	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	5	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	6	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	7	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	8	recorded	record	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	9	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	10	high	high	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	12	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	13	important	important	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	14	issue	issue	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	15	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	17	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	18	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	19	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	20	located	locate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	21	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	23	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	24	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	25	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	26	reduce	reduce	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	28	resource	resource	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	29	consump-	consump-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	30	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	169	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	2	deal	deal	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	3	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	5	concern	concern	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	7	Yang	Yang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	8	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	9	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	170	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	2	49	49	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	4	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	6	data	data	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	7	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	8	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	9	record	record	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	10	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	11	sessions	session	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	13	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	14	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	15	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	16	possible	possible	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	17	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	18	apply	apply	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	19	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	20	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	21	mining	mining	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	22	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	171	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	2	proposed	proposed	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	3	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	4	works	work	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	6	two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	7	steps	step	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	8	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	9	first	first	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	11	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	12	sessions	session	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	13	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	14	clustered	cluster	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	15	;	;	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	16	second	second	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	18	transition	transition	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	19	probabilities	probability	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	20	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	21	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	22	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	24	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	25	cluster	cluster	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	26	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	27	computed	compute	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	172	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	1	7	7	LS
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	2	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	3	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	4	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	5	way	way	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	8	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	9	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	10	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	11	Pallis	Pallis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	12	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	13	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	173	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	2	37	37	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	4	identifies	identify	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	5	clusters	cluster	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	7	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	8	whose	whose	WP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	9	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	10	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	11	correlated	correlate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	12	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	13	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	14	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	15	step	step	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	16	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	17	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	18	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	174	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	1	Other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	2	approaches	approach	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	3	take	take	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	4	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	5	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	7	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	9	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	10	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	11	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	12	points	point	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	14	view	view	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	15	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	16	complement	complement	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	18	sequences	sequence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	19	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	20	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	21	one	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	22	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	175	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	3	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	5	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	6	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	7	hit	hit	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	10	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	11	cache	cache	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	12	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	13	hidden	hide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	14	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	16	origin	origin	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	17	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	176	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	2	fact	fact	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	3	distorts	distort	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	5	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	7	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	8	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	10	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	12	prevents	prevent	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	14	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	15	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	16	making	make	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	17	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	18	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	19	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	20	that	that	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	21	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	177	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	178	1	Chen	Chen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	178	2	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	178	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	178	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	2	9	9	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	4	deal	deal	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	5	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	6	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	7	problem	problem	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	8	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	9	proposing	propose	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	10	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	11	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	12	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	13	coordinates	coordinate	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	14	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	15	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	16	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	17	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	18	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	19	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	179	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	1	Another	another	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	2	coordinated	coordinated	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	3	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	5	taken	take	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	6	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	7	Pons	Pons	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	8	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	9	39	39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	10	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	11	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	13	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	15	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	17	application	application	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	19	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	20	considering	consider	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	21	coordination	coordination	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	22	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	23	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	24	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	25	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	180	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	3	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	6	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	7	builds	build	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	9	generic	generic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	10	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	11	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	12	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	13	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	14	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	15	receives	receive	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	17	adapts	adapt	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	18	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	19	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	20	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	21	own	own	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	22	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	181	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	1	Some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	2	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	3	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	4	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	5	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	7	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	9	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	10	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	12	only	only	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	13	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	15	but	but	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	16	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	17	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	18	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	19	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	20	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	21	work	work	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	22	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	23	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	24	specific	specific	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	25	element	element	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	28	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	182	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	2	improve	improve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	3	PPM	PPM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	5	accuracy	accuracy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	7	Ban	Ban	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	8	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	9	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	183	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	2	3	3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	4	propose	propose	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	6	include	include	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	7	extra	extra	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	8	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	9	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	11	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	12	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	13	describe	describe	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	14	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	15	node	node	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	16	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	17	capability	capability	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	184	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	1	Chen	Chen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	3	Zhang	Zhang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	4	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	5	8	8	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	6	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	7	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	9	memory	memory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	11	efficient	efficient	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	12	version	version	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	14	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	15	PPM	ppm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	16	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	17	whose	whose	WP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	18	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	19	tree	tree	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	20	branches	branch	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	21	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	22	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	23	height	height	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	24	depending	depend	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	25	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	27	root	root	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	28	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	29	popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	185	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	2	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	5	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	6	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	7	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	9	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	10	orders	order	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	12	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	13	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	14	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	15	Dongshan	Dongshan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	17	Junyi	Junyi	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	18	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	19	20	20	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	20	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	21	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	22	deal	deal	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	23	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	25	scalability	scalability	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	28	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	29	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	186	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	2	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	3	presented	present	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	4	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	5	Gunduz	Gunduz	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	7	Ozsu	Ozsu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	8	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	9	24	24	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	10	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	11	compare	compare	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	12	session	session	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	13	similarities	similarity	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	14	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	15	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	16	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	187	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	1	Sessions	session	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	2	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	3	modeled	model	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	4	taking	take	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	5	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	6	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	8	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	10	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	11	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	14	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	15	spent	spend	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	17	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	18	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	188	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	2	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	3	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	4	Bonino	Bonino	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	5	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	6	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	189	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	2	7	7	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	4	describes	describe	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	5	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	6	initial	initial	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	7	set	set	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	9	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	10	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	11	repre-	repre-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	12	sented	sente	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	13	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	14	finite	finite	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	15	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	16	state	state	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	17	machines	machine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	18	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	19	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	20	evolve	evolve	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	21	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	22	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	23	genetic	genetic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	24	operators	operator	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	190	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	1	One	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	3	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	4	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	6	selected	select	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	7	every	every	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	8	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	10	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	11	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	12	received	receive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	13	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	14	order	order	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	15	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	16	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	17	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	18	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	191	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	2	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	3	representing	representing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	4	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	5	sessions	session	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	7	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	8	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	9	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	10	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	11	employed	employ	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	13	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	14	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	16	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	17	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	18	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	19	represented	represent	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	20	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	21	edges	edge	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	24	nodes	node	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	25	represent	represent	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	26	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	27	instead	instead	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	28	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	29	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	192	30	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	1	3.2.2	3.2.2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	2	Hybrid	hybrid	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	3	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	4	Predictors	Predictors	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	5	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	6	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	7	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	8	subcategory	subcategory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	9	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	10	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	11	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	12	complement	complement	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	14	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	16	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	17	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	18	another	another	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	19	source	source	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	21	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	23	e.g.	e.g.	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	26	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	27	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	28	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	29	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	30	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	193	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	2	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	3	reasons	reason	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	4	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	5	take	take	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	6	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	7	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	8	extra	extra	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	9	sources	source	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	10	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	11	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	12	either	either	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	14	increase	increase	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	16	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	17	accuracy	accuracy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	18	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	19	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	20	control	control	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	22	computational	computational	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	23	complexity	complexity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	26	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	194	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	1	Three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	5	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	6	analyzed	analyze	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	7	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	8	Zukerman	Zukerman	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	9	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	10	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	195	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	2	50	50	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	4	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	7	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	8	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	9	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	10	employ	employ	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	12	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	14	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	15	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	17	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	18	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	20	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	21	source	source	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	196	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	1	One	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	3	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	6	Linked	Linked	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	7	Space	Space	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	9	Time	Time	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	10	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	11	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	13	com-	com-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	14	bines	bin	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	16	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	18	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	19	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	21	referrer	referrer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	22	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	23	found	find	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	24	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	26	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	27	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	197	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	2	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	3	later	later	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	4	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	5	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	6	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	7	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	9	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	10	combine	combine	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	11	all	all	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	13	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	14	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	15	build	build	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	16	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	17	single	single	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	18	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	198	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	2	Double	Double	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	3	Dependency	Dependency	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	4	Graph	Graph	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	5	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	6	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	7	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	9	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	10	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	11	see	see	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	12	Section	Section	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	13	3.3	3.3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	14	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	15	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	16	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	17	Domenech	Domenech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	18	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	19	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	199	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	2	18	18	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	4	employs	employ	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	6	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	8	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	9	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	10	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	11	build	build	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	12	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	13	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	14	dependency	dependency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	15	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	16	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	17	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	19	DG	DG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	20	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	21	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	22	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	23	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	24	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	25	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	200	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	2	predictor	predictor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	3	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	4	requires	require	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	6	analyze	analyze	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	7	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	8	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	10	distinguish	distinguish	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	11	container	container	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	12	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	13	embedded	embed	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	14	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	201	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	2	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	3	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	4	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	5	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	6	Georgakis	Georgakis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	8	Li	Li	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	9	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	10	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	11	23	23	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	12	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	13	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	14	described	describe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	16	Section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	17	3.4	3.4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	18	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	19	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	20	takes	take	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	21	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	22	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	24	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	26	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	27	received	receive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	28	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	29	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	30	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	202	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	1	8	8	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	2	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	4	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	6	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	7	interested	interested	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	10	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	12	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	13	read	read	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	14	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	17	path	path	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	18	followed	follow	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	19	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	21	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	203	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	204	1	Nanopoulos	Nanopoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	204	2	et	et	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	204	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	204	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	2	35	35	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	4	propose	propose	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	5	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	6	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	7	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	8	builds	build	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	9	association	association	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	10	rules	rule	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	11	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	13	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	14	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	15	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	17	allowing	allow	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	18	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	19	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	20	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	21	transitions	transition	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	22	found	find	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	23	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	25	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	26	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	29	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	30	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	205	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	2	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	3	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	4	way	way	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	7	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	8	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	9	Mabroukeh	Mabroukeh	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	11	Ezeife	Ezeife	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	12	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	13	32	32	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	14	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	15	builds	build	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	16	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	17	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	18	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	19	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	20	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	21	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	22	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	206	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	3	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	5	case	case	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	9	tree	tree	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	10	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	11	pruned	prune	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	12	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	13	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	15	concept	concept	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	17	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	18	distance	distance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	207	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	2	compute	compute	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	4	distance	distance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	7	domain	domain	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	8	ontology	ontology	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	10	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	11	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	12	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	13	during	during	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	15	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	18	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	19	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	21	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	22	required	require	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	208	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	1	3.3	3.3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	2	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	3	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	4	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	5	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	6	includes	include	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	7	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	9	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	11	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	13	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	14	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	15	either	either	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	16	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	18	only	only	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	19	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	20	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	21	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	22	combination	combination	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	23	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	24	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	25	inputs	input	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	209	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	2	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	3	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	4	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	5	issued	issue	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	6	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	7	sending	send	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	8	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	9	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	11	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	12	contain	contain	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	13	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	14	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	15	providing	provide	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	16	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	17	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	18	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	19	details	detail	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	21	besides	besides	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	23	URI	URI	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	210	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	1	One	one	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	4	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	5	used	used	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	6	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	7	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	8	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	10	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	11	Referer	Referer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	12	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	211	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	1	Although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	3	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	4	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	5	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	6	optional	optional	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	7	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	9	standard	standard	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	10	protocol	protocol	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	13	most	most	RBS
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	14	popular	popular	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	15	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	16	browsers	browser	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	17	always	always	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	18	provide	provide	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	19	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	20	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	21	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	22	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	212	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	2	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	4	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	5	discover	discover	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	6	first-	first-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	7	order	order	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	8	dependences	dependence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	9	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	10	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	11	observing	observe	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	12	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	13	Requested	Requested	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	14	Object	Object	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	15	URI	URI	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	17	Referrer	Referrer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	18	URI	URI	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	19	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	20	pairs	pair	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	213	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	1	Together	together	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	2	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	4	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	5	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	9	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	10	returns	return	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	11	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	12	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	13	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	14	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	214	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	1	By	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	2	looking	look	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	3	at	at	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	5	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	6	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	9	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	10	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	11	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	12	take	take	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	13	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	14	features	feature	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	17	requested	requested	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	18	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	19	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	20	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	215	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	1	Response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	2	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	3	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	5	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	6	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	7	include	include	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	8	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	9	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	11	Type	type	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	12	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	14	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	15	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	16	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	17	Length	Length	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	18	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	19	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	216	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	217	1	Zukerman	Zukerman	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	217	2	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	217	3	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	217	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	2	50	50	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	4	compare	compare	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	5	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	6	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	7	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	8	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	9	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	10	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	11	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	12	take	take	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	13	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	14	account	account	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	15	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	16	factors	factor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	17	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	19	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	21	documents	document	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	22	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	25	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	26	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	27	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	28	inferred	infer	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	29	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	30	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	31	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	32	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	33	Referer	Referer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	34	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	35	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	36	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	37	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	38	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	39	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	40	combina-	combina-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	41	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	42	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	43	both	both	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	218	44	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	1	Their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	2	results	result	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	3	show	show	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	4	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	6	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	7	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	8	combines	combine	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	9	all	all	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	10	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	11	factors	factor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	14	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	15	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	16	yields	yield	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	18	best	good	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	19	predictive	predictive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	20	power	power	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	22	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	23	terms	term	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	25	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	26	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	27	space	space	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	28	consumption	consumption	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	219	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	2	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	3	authors	author	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	4	propose	propose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	7	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	8	later	later	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	9	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	10	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	11	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	12	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	15	max-	max-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	16	Hybrid	Hybrid	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	17	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	18	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	20	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	21	uses	use	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	22	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	23	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	25	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	26	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	27	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	28	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	29	timestamp	timestamp	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	30	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	31	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	32	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	33	Referer	Referer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	34	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	35	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	36	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	37	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	38	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	39	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	40	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	41	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	42	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	43	Length	Length	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	44	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	45	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	46	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	220	47	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	2	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	3	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	4	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	5	18	18	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	6	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	8	classified	classify	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	10	introduced	introduce	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	11	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	13	previous	previous	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	14	section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	16	employs	employ	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	18	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	19	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	20	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	21	Type	Type	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	22	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	23	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	24	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	25	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	26	distinguish	distinguish	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	27	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	28	container	container	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	29	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	30	embedded	embed	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	31	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	221	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	1	Container	container	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	2	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	4	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	5	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	7	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	8	demand	demand	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	9	ex-	ex-	XX
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	10	plicitly	plicitly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	12	while	while	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	13	secondary	secondary	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	14	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	15	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	16	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	17	embedded	embed	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	18	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	20	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	22	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	23	automatically	automatically	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	24	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	26	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	27	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	222	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	2	allows	allow	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	3	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	4	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	5	differentiate	differentiate	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	7	depen-	depen-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	8	dences	dence	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	9	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	10	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	13	same	same	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	14	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	15	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	16	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	17	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	18	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	19	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	223	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	1	Thus	thus	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	3	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	4	represents	represent	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	6	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	10	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	11	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	12	faithfully	faithfully	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	13	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	14	DG	DG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	224	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	2	Referrer	Referrer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	3	Graph	Graph	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	4	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	5	RG	RG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	6	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	7	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	8	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	9	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	10	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	11	Ossa	Ossa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	12	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	13	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	225	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	2	16	16	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	7	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	226	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	1	Both	both	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	3	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	5	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	6	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	7	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	8	Type	Type	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	9	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	10	response	response	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	11	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	12	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	13	classify	classify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	14	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	15	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	16	primary	primary	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	17	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	18	container	container	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	19	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	20	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	21	secondary	secondary	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	22	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	23	embedded	embedded	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	24	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	25	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	227	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	1	However	however	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	3	unlike	unlike	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	4	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	6	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	7	uses	use	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	9	temporal	temporal	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	10	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	12	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	14	RG	RG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	15	uses	use	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	17	URI	uri	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	20	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	21	Referrer	Referrer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	22	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	23	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	24	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	25	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	26	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	27	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	28	individual	individual	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	29	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	30	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	31	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	32	build	build	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	33	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	34	arcs	arc	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	35	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	36	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	37	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	228	38	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	3	way	way	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	5	RG	RG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	6	builds	build	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	7	arcs	arc	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	8	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	9	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	10	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	11	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	12	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	13	really	really	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	14	linked	link	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	15	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	16	hyper	hyper	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	17	references	reference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	18	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	20	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	21	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	24	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	25	9	9	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	26	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	27	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	28	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	29	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	30	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	31	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	32	sequentially	sequentially	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	33	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	34	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	35	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	36	browser	browser	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	229	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	2	allows	allow	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	3	RG	RG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	4	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	5	represent	represent	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	7	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	9	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	10	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	11	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	12	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	13	faithfully	faithfully	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	14	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	15	DDG	DDG	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	230	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	1	HTTP	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	2	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	3	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	4	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	5	considered	consider	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	6	accurate	accurate	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	7	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	8	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	9	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	10	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	11	explicitly	explicitly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	12	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	14	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	15	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	17	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	231	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	2	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	3	words	word	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	5	there	there	EX
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	6	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	7	no	no	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	8	derived	derive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	9	guessing	guessing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	11	might	may	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	12	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	13	erroneously	erroneously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	14	interpreted	interpret	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	15	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	16	happens	happen	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	17	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	19	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	21	requets	requet	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	22	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	232	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	2	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	4	unnecessary	unnecessary	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	6	access	access	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	8	full	full	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	9	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	10	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	12	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	13	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	14	allow	allow	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	15	shorter	short	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	16	processing	processing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	17	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	19	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	20	less	less	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	21	intrusive	intrusive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	22	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	23	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	24	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	25	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	26	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	27	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	233	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	2	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	3	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	4	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	5	continually	continually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	8	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	9	allows	allow	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	11	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	12	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	13	actively	actively	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	14	capture	capture	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	15	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	16	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	234	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	1	3.4	3.4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	2	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	3	Prediction	Prediction	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	4	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	5	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	6	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	8	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	9	extracted	extract	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	10	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	12	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	14	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	15	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	16	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	17	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	18	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	19	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	20	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	21	included	include	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	22	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	23	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	24	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	235	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	1	These	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	2	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	3	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	4	use	use	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	6	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	8	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	10	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	11	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	13	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	14	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	16	usually	usually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	17	without	without	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	18	requiring	require	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	19	history	history	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	20	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	236	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	3	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	4	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	5	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	6	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	7	specially	specially	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	8	designed	design	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	10	deal	deal	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	11	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	12	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	13	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	14	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	15	newly	newly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	16	created	create	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	17	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	18	never	never	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	19	visited	visit	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	20	before	before	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	22	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	23	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	24	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	25	sites	site	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	26	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	27	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	28	URL	url	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	29	could	could	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	30	have	have	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	31	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	32	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	33	dependent	dependent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	34	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	237	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	1	Despite	despite	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	3	wide	wide	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	4	variety	variety	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	6	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	7	falling	fall	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	8	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	9	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	10	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	12	we	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	13	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	14	identified	identify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	15	three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	16	trends	trend	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	17	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	18	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	19	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	20	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	21	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	22	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	23	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	24	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	238	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	1	3.4.1	3.4.1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	2	Using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	3	hyperlinks	hyperlink	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	4	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	5	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	6	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	7	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	8	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	9	used	use	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	11	references	reference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	12	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	13	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	14	found	find	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	15	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	16	parsing	parse	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	18	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	19	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	20	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	22	without	without	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	23	any	any	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	24	further	further	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	25	analysis	analysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	239	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	1	Chinen	Chinen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	3	Ya-	Ya-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	4	maguchi	maguchi	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	5	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	6	10	10	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	7	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	8	designed	design	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	10	implemented	implement	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	11	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	12	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	13	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	14	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	15	called	call	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	17	WWW	WWW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	18	Collector	Collector	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	19	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	20	Wcol	Wcol	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	21	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	22	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	23	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	24	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	25	idea	idea	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	240	26	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	2	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	3	prefetches	prefetche	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	5	stores	store	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	6	referenced	reference	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	7	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	10	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	11	shares	share	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	12	prefetched	prefetche	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	13	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	14	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	15	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	16	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	17	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	241	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	1	Cohen	Cohen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	3	Kaplan	Kaplan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	4	describe	describe	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	6	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	7	12	12	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	8	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	9	three	three	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	10	simple	simple	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	11	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	12	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	13	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	14	pre-	pre-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	15	resolving	resolve	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	16	host	host	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	17	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	18	names	name	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	20	pre	pre	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	21	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	22	connecting	connecting	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	24	pre	pre	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	25	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	26	warming	warming	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	27	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	28	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	29	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	30	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	31	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	32	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	33	non	non	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	34	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	35	prefetchable	prefetchable	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	36	URLs	url	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	242	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	1	Predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	2	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	3	performed	perform	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	4	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	5	extracting	extract	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	6	URLs	url	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	7	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	8	anchor	anchor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	9	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	11	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	12	visited	visit	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	13	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	243	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	2	study	study	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	3	considers	consider	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	4	two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	5	classes	class	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	7	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	244	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	2	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	3	class	class	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	4	includes	include	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	5	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	6	linked	link	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	7	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	9	results	result	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	10	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	12	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	13	search	search	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	14	engine	engine	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	15	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	17	portal	portal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	19	whereas	whereas	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	21	second	second	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	22	class	class	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	23	contains	contain	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	24	all	all	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	25	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	26	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	27	were	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	28	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	29	preceded	precede	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	30	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	31	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	32	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	33	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	34	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	35	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	36	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	37	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	38	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	39	60	60	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	40	seconds	second	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	245	41	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	1	More	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	2	recently	recently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	6	context	context	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	8	CDNs	cdn	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	10	Sidiropoulos	Sidiropoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	11	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	12	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	246	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	2	44	44	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	4	identify	identify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	5	clus-	clus-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	6	ters	ter	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	8	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	9	correlated	correlate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	10	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	11	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	12	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	13	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	14	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	15	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	16	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	17	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	18	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	19	communities	community	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	20	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	22	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	23	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	24	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	25	communities	community	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	27	basic	basic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	28	outsourcing	outsourcing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	29	unit	unit	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	30	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	31	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	32	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	33	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	34	prefetched	prefetche	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	247	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	2	ap-	ap-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	3	proach	proach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	6	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	9	hyperlink	hyperlink	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	10	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	12	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	13	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	14	assume	assume	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	15	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	16	two	two	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	17	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	18	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	19	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	20	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	21	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	22	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	23	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	24	correlated	correlate	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	25	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	26	if	if	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	27	there	there	EX
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	28	exists	exist	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	29	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	30	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	31	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	32	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	248	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	3	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	4	Nanopoulos	Nanopoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	5	et	et	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	6	al	al	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	249	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	2	35	35	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	5	described	describe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	7	Section	section	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	8	3.2	3.2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	10	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	11	falls	fall	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	12	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	13	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	14	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	15	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	16	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	17	complements	complement	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	19	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	20	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	21	obtained	obtain	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	22	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	24	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	26	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	27	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	29	hyperlink	hyperlink	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	30	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	31	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	33	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	34	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	250	35	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	2	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	3	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	4	applied	apply	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	5	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	7	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	8	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	10	allow	allow	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	11	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	12	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	13	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	14	transitions	transition	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	15	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	16	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	17	included	include	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	18	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	19	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	251	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	1	3.4.2	3.4.2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	2	Semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	3	approaches	approach	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	4	Algorithms	Algorithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	6	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	7	subcategory	subcategory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	8	tend	tend	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	9	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	10	capture	capture	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	12	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	13	surfing	surf	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	14	interest	interest	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	15	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	16	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	17	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	19	visited	visit	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	20	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	252	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	1	10	10	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	2	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	3	keyword	keyword	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	4	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	6	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	7	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	8	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	9	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	10	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	11	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	12	Xu	Xu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	14	Ibrahim	Ibrahim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	15	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	16	48	48	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	17	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	253	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	3	lies	lie	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	4	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	6	fact	fact	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	7	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	8	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	9	surfing	surfing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	10	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	11	often	often	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	12	guided	guide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	13	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	14	some	some	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	15	keywords	keyword	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	18	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	19	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	20	surrounds	surround	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	21	hyperlink	hyperlink	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	22	definitions	definition	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	23	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	24	hrefs	hrefs	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	25	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	26	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	27	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	28	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	254	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	1	Therefore	therefore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	4	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	5	predicts	predict	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	6	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	7	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	8	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	10	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	11	preferences	preference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	13	past	past	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	14	retrieved	retrieve	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	15	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	255	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	2	do	do	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	3	so	so	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	6	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	7	ranks	rank	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	9	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	12	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	13	visited	visited	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	14	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	15	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	16	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	18	similarity	similarity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	21	keywords	keyword	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	22	found	find	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	23	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	24	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	25	“	"	``
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	26	anchor	anchor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	27	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	28	”	"	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	29	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	30	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	31	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	33	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	34	accessed	access	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	35	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	256	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	1	More	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	2	recently	recently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	4	Georgakis	Georgakis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	6	Li	Li	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	7	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	8	23	23	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	9	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	10	propose	propose	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	11	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	12	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	13	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	14	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	15	considers	consider	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	17	anchored	anchor	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	18	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	19	around	around	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	20	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	23	outbound	outbound	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	24	hyperlinks	hyperlinks	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	257	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	1	To	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	2	assign	assign	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	3	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	4	weight	weight	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	5	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	6	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	7	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	8	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	9	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	12	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	13	keeps	keep	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	14	counters	counter	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	17	frequencies	frequency	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	19	appearance	appearance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	21	bigrams	bigram	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	22	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	23	visited	visited	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	24	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	25	non	non	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	26	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	27	visited	visited	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	28	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	258	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	1	Davison	Davison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	3	14	14	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	5	proposes	propose	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	6	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	7	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	8	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	9	predicts	predict	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	11	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	12	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	13	next	next	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	14	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	15	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	16	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	17	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	18	conducting	conduct	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	19	an	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	20	analysis	analysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	23	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	24	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	25	his	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	26	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	27	her	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	28	recently	recently	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	29	requested	request	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	30	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	31	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	259	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	2	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	3	compares	compare	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	5	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	6	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	8	around	around	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	10	hypertext	hypertext	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	11	anchors	anchor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	14	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	15	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	16	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	18	text	text	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	19	contained	contain	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	20	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	21	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	22	visited	visit	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	23	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	260	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	1	3.4.3	3.4.3	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	2	Labeled	label	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	3	approaches	approach	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	4	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	5	subcategory	subcategory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	6	includes	include	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	7	all	all	PDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	8	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	9	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	10	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	12	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	13	designer	designer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	14	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	15	in-	in-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	16	volved	volve	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	17	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	19	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	20	process	process	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	21	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	22	including	include	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	23	extra	extra	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	24	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	25	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	26	will	will	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	27	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	28	later	later	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	29	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	30	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	31	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	32	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	33	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	261	34	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	2	common	common	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	3	advantage	advantage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	5	these	these	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	6	specific	specific	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	8	designer	designer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	9	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	10	dependent	dependent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	11	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	12	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	13	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	14	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	15	exhibit	exhibit	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	16	interesting	interesting	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	17	results	result	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	18	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	20	environments	environment	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	21	where	where	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	22	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	23	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	24	tested	test	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	25	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	26	that	that	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	27	is	is	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	28	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	29	those	those	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	30	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	31	which	which	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	32	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	33	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	34	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	35	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	36	developed	develop	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	262	37	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	2	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	3	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	4	their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	5	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	6	disadvantage	disadvantage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	8	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	9	they	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	10	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	11	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	12	generic	generic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	13	solutions	solution	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	14	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	15	could	could	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	16	be	be	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	17	easily	easily	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	18	adapted	adapt	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	19	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	20	general	general	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	22	real	real	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	23	scenarios	scenario	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	263	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	1	Khan	Khan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	2	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	3	Tao	Tao	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	4	identify	identify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	5	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	6	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	7	27	27	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	8	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	9	four	four	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	10	dominant	dominant	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	11	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	13	hyperlink	hyperlink	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	14	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	15	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	16	Chain	Chain	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	18	Tree	Tree	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	20	Complete	Complete	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	21	Graph	Graph	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	24	Tree	tree	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	25	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	26	Complete	complete	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	27	Core	Core	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	28	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	264	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	1	Their	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	2	pro-	pro-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	3	posal	posal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	4	consists	consist	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	5	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	6	labeling	label	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	7	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	8	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	9	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	10	kind	kind	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	12	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	14	develop	develop	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	15	smarter	smart	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	16	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	17	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	18	when	when	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	20	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	22	webspace	webspace	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	23	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	24	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	25	brought	bring	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	26	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	27	con-	con-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	28	sideration	sideration	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	265	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	1	Pons	pon	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	3	40	40	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	5	proposes	propose	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	6	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	7	conduct	conduct	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	9	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	10	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	11	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	12	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	13	links	link	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	266	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	1	That	that	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	2	is	is	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	4	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	5	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	6	explicitly	explicitly	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	7	embedded	embed	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	8	during	during	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	10	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	11	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	13	associ-	associ-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	14	ated	ate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	15	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	16	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	17	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	18	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	19	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	267	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	2	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	3	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	5	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	6	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	7	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	8	hyperlink	hyperlink	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	10	since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	12	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	13	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	14	represents	represent	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	15	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	16	pointer	pointer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	17	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	18	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	19	type	type	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	20	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	21	meaning	meaning	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	22	directed	direct	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	23	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	24	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	25	predecessor	predecessor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	26	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	27	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	28	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	29	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	30	successor	successor	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	31	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	32	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	268	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	1	Examples	example	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	3	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	4	link	link	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	5	types	type	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	6	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	7	sequential	sequential	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	9	similar	similar	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	11	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	13	cause	cause	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	14	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	15	effective	effective	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	17	implication	implication	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	19	etc	etc	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	269	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	2	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	4	Mabroukeh	Mabroukeh	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	5	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	6	32	32	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	7	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	8	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	9	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	10	previously	previously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	11	classified	classify	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	12	into	into	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	13	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	14	group	group	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	16	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	17	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	19	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	20	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	21	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	22	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	23	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	24	because	because	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	25	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	26	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	27	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	28	main	main	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	29	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	30	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	31	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	32	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	270	33	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	1	Nevertheless	nevertheless	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	3	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	4	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	5	actually	actually	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	6	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	7	hybrid	hybrid	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	8	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	9	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	10	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	11	employs	employ	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	12	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	13	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	14	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	15	prune	prune	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	17	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	18	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	19	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	20	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	21	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	22	computed	computed	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	23	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	24	distance	distance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	271	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	2	proposal	proposal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	3	assumes	assume	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	4	that	that	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	6	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	7	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	8	is	be	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	9	available	available	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	10	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	12	form	form	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	14	domain	domain	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	15	ontology	ontology	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	17	provided	provide	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	18	during	during	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	20	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	23	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	24	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	272	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	1	4	4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	2	Conclusions	Conclusions	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	3	Since	since	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	5	inception	inception	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	8	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	11	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	13	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	14	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	15	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	16	adapted	adapt	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	17	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	18	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	19	increasing	increase	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	20	complexity	complexity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	22	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	23	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	24	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	273	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	1	Thus	thus	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	4	com-	com-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	5	plexity	plexity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	8	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	9	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	10	increased	increase	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	12	too	too	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	274	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	3	paper	paper	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	5	we	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	6	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	7	summarized	summarize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	8	11	11	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	9	Table	table	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	10	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	11	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	12	Surveyed	survey	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	13	papers	paper	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	14	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	15	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	17	year	year	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	18	CATEGORY	CATEGORY	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	19	1995	1995	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	20	-	-	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	21	2000	2000	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	22	2001	2001	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	23	-	-	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	24	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	25	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	26	-	-	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	27	2010	2010	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	28	Obj	obj	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	275	29	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	1	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	3	6	6	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	5	5	5	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	7	33	33	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	9	21	21	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	10	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	11	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	12	26	26	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	14	46	46	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	16	30	30	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	18	42	42	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	19	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	20	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	21	47	47	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	22	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	23	Seq	seq	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	276	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	1	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	2	pure	pure	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	3	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	4	4	4	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	6	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	8	43	43	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	10	13	13	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	12	22	22	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	14	38	38	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	15	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	16	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	17	8	8	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	19	20	20	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	21	7	7	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	23	24	24	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	25	28	28	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	26	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	27	49	49	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	28	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	29	9	9	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	30	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	31	25	25	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	32	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	33	39	39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	34	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	35	45	45	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	36	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	37	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	38	31	31	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	39	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	40	3	3	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	41	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	42	37	37	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	43	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	44	hybrid	hybrid	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	45	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	46	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	47	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	48	50	50	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	49	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	50	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	51	35	35	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	52	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	53	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	54	23	23	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	55	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	56	32	32	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	57	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	58	18	18	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	59	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	60	HTTP	HTTP	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	61	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	62	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	63	2	2	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	64	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	65	50	50	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	66	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	67	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	68	16	16	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	69	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	70	18	18	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	71	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	72	Content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	73	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	74	10	10	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	75	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	76	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	77	12	12	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	78	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	79	14	14	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	80	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	81	27	27	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	82	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	83	35	35	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	84	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	85	48	48	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	86	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	87	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	88	23	23	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	89	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	90	40	40	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	91	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	92	44	44	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	93	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	94	32	32	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	95	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	96	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	97	categorized	categorize	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	98	more	more	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	99	than	than	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	100	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	101	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	102	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	103	research	research	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	104	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	105	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	106	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	107	ap-	ap-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	108	plied	ply	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	109	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	110	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	111	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	112	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	277	113	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	1	To	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	2	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	3	end	end	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	5	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	6	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	7	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	8	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	11	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	12	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	13	used	use	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	14	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	15	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	16	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	17	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	18	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	19	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	20	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	21	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	22	successfully	successfully	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	23	applied	apply	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	278	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	1	Through	through	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	3	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	4	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	7	pros	pro	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	9	cons	con	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	10	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	11	each	each	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	12	category	category	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	14	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	15	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	16	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	17	discussed	discuss	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	279	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	1	Table	table	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	2	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	3	summarizes	summarize	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	5	classification	classification	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	8	surveyed	survey	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	9	papers	paper	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	10	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	11	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	13	taxonomy	taxonomy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	14	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	15	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	16	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	17	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	19	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	20	year	year	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	21	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	22	publication	publication	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	280	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	1	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	2	table	table	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	3	permits	permit	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	4	us	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	6	identify	identify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	7	trends	trend	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	9	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	10	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	281	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	2	one	one	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	3	can	can	MD
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	4	observe	observe	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	6	dur-	dur-	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	7	ing	e	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	9	first	first	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	10	period	period	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	12	five	five	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	13	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	15	most	most	JJS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	16	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	18	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	19	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	20	dealt	deal	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	21	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	22	with	with	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	23	object	object	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	24	characteristics	characteristic	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	25	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	26	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	27	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	28	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	29	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	30	make	make	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	31	predictions	prediction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	282	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	3	next	next	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	4	five	five	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	5	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	6	year	year	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	7	period	period	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	10	predominant	predominant	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	11	choice	choice	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	12	was	be	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	13	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	14	use	use	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	15	only	only	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	17	sequence	sequence	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	19	accesses	access	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	20	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	21	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	22	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	23	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	24	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	25	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	26	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	27	although	although	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	28	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	29	number	number	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	30	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	31	content-	content-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	32	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	33	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	34	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	35	increased	increase	VBD
work_jsnmrvhjlfgxfkyxshfjgxeayu	283	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	1	Finally	finally	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	3	during	during	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	5	last	last	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	6	five	five	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	7	years	year	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	9	researchers	researcher	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	10	have	have	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	11	moved	move	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	12	towards	towards	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	13	more	more	RBR
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	14	sophisticated	sophisticated	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	15	inputs	input	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	284	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	1	Algorithms	Algorithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	2	considering	consider	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	3	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	4	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	5	types	type	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	6	as	as	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	7	well	well	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	8	as	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	9	predictors	predictor	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	10	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	11	analyze	analyze	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	12	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	13	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	14	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	15	or	or	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	16	require	require	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	17	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	18	collaboration	collaboration	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	20	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	21	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	22	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	23	designer	designer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	24	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	25	gaining	gain	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	26	relative	relative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	27	importance	importance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	28	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	29	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	30	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	31	trend	trend	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	285	32	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	1	Furthermore	furthermore	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	3	it	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	4	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	5	also	also	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	6	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	7	observed	observe	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	9	increasing	increase	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	10	concern	concern	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	11	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	12	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	13	computational	computational	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	14	costs	cost	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	15	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	16	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	17	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	286	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	1	Despite	despite	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	3	this	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	5	there	there	EX
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	6	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	7	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	8	gaps	gap	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	9	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	10	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	11	design	design	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	12	space	space	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	14	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	15	algo-	algo-	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	16	rithms	rithms	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	17	that	that	WDT
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	18	remain	remain	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	19	unexplored	unexplored	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	287	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	1	Few	few	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	2	papers	paper	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	4	i.e.	i.e.	FW
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	6	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	7	21	21	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	9	9	9	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	11	39	39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	12	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	14	analyze	analyze	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	15	several	several	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	17	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	18	points	point	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	19	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	20	view	view	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	21	simultaneously	simultaneously	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	23	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	24	none	none	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	25	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	26	them	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	27	combine	combine	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	28	different	different	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	29	input	input	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	30	categories	category	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	288	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	1	Besides	besides	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	3	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	4	use	use	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	6	HTTP	HTTP	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	7	headers	header	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	8	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	9	not	not	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	10	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	11	intensively	intensively	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	12	explored	explore	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	14	no	no	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	15	algorithm	algorithm	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	16	combining	combine	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	17	header	header	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	19	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	20	analysis	analysis	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	21	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	22	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	23	proposed	propose	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	289	24	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	1	As	as	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	2	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	3	future	future	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	4	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	6	we	-PRON-	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	7	plan	plan	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	8	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	9	classify	classify	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	10	other	other	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	11	aspects	aspect	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	12	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	13	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	14	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	15	systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	16	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	17	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	18	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	19	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	20	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	21	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	22	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	23	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	24	location	location	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	25	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	26	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	27	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	28	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	29	;	;	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	30	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	31	proposals	proposal	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	32	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	33	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	34	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	35	architectures	architecture	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	36	according	accord	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	37	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	38	how	how	WRB
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	39	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	40	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	41	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	42	engines	engine	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	43	are	be	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	44	coordinated	coordinate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	45	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	46	thus	thus	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	47	providing	provide	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	48	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	49	complete	complete	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	50	study	study	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	51	about	about	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	52	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	53	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	54	techniques	technique	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	55	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	56	its	-PRON-	PRP$
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	57	beginnings	beginning	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	58	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	59	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	60	days	day	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	290	61	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	1	Acknowledgments	acknowledgment	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	2	This	this	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	3	work	work	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	4	has	have	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	5	been	be	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	6	partially	partially	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	7	supported	support	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	8	by	by	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	9	Spanish	Spanish	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	10	Ministry	Ministry	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	12	Science	Science	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	14	Innovation	Innovation	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	15	under	under	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	16	grant	grant	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	17	TIN2009	TIN2009	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	18	-	-	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	19	08201	08201	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	21	Generalitat	Generalitat	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	22	Valenciana	Valenciana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	23	under	under	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	24	grant	grant	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	25	GV/2011/002	GV/2011/002	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	26	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	27	Universitat	Universitat	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	28	Politecnica	Politecnica	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	29	de	de	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	30	Valencia	Valencia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	31	under	under	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	32	grant	grant	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	33	PAID-06-	PAID-06-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	34	12	12	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	35	10/2424	10/2424	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	291	36	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	1	References	reference	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	3	1	1	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	5	Silvana	Silvana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	6	Vanesa	Vanesa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	7	Aciar	Aciar	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	9	Christian	Christian	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	10	Serarols	Serarols	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	11	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	12	Tarres	Tarres	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	14	Marcelo	Marcelo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	15	Royo	Royo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	16	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	17	Vela	Vela	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	19	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	20	Josep	Josep	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	21	Lluis	Lluis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	22	de	de	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	23	la	la	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	24	Rosa	Rosa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	25	i	i	PRP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	26	Esteva	Esteva	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	292	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	1	Increasing	increase	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	2	effectiveness	effectiveness	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	3	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	4	e	e	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	5	-	-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	6	commerce	commerce	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	7	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	8	rec-	rec-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	9	ommendations	ommendation	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	10	applying	apply	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	11	intelligent	intelligent	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	12	agents	agent	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	293	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	1	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	2	Journal	Journal	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	4	Busi-	Busi-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	5	ness	ness	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	7	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	8	Research	Research	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	10	1(1):81–97	1(1):81–97	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	12	2007	2007	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	294	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	2	2	2	LS
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	4	David	David	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	5	W.	W.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	6	Albrecht	Albrecht	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	8	Ingrid	Ingrid	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	9	Zukerman	Zukerman	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	12	Ann	Ann	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	13	E.	E.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	14	Nicholson	Nicholson	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	295	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	1	Pre	pre	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	2	-	-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	3	sending	send	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	4	documents	document	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	5	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	6	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	7	WWW	WWW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	8	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	9	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	10	comparative	comparative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	11	study	study	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	296	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	5	Six-	Six-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	6	teenth	teenth	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	7	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	8	Joint	Joint	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	9	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	11	Artificial	Artificial	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	12	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	14	Stockholm	Stockholm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	16	Sweden	Sweden	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	18	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	297	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	2	3	3	LS
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	4	Zhijie	Zhijie	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	5	Ban	Ban	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	7	Zhimin	Zhimin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	8	Gu	Gu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	11	Yu	Yu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	12	Jin	Jin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	298	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	2	ppm	ppm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	3	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	4	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	5	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	7	stochastic	stochastic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	8	gradient	gradient	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	9	descent	descent	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	10	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	11	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	12	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	299	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	6	22nd	22nd	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	7	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	8	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	10	Advanced	Advanced	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	11	Information	Information	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	12	Networking	Networking	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	14	Applications	Applications	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	15	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	16	AINA	AINA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	17	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	19	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	20	166–173	166–173	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	22	Okinawa	Okinawa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	24	Japan	Japan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	25	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	26	2008	2008	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	300	27	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	2	4	4	LS
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	4	Azer	Azer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	5	Bestavros	Bestavros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	301	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	1	Using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	2	speculation	speculation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	3	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	4	reduce	reduce	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	5	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	6	load	load	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	8	service	service	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	9	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	11	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	12	WWW	WWW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	302	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	5	4th	4th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	6	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	7	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	8	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	10	Information	Information	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	12	Knowledge	Knowledge	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	13	Management	Management	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	15	Baltimore	Baltimore	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	17	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	19	1995	1995	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	303	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	2	5	5	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	4	Azer	Azer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	5	Bestavros	Bestavros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	304	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	1	Speculative	speculative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	2	data	data	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	3	dissemination	dissemination	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	5	service	service	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	6	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	7	reduce	reduce	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	8	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	9	load	load	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	11	network	network	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	12	traffic	traffic	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	14	service	service	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	15	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	16	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	17	distributed	distribute	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	18	information	information	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	19	systems	system	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	305	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	5	Twelfth	Twelfth	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	7	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	9	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	10	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	12	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	13	180–187	180–187	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	15	New	New	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	16	Orleans	Orleans	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	18	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	20	1996	1996	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	306	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	2	6	6	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	4	Azer	Azer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	5	Bestavros	Bestavros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	7	C.	C.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	8	Cunha	Cunha	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	307	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	1	Server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	3	initiated	initiate	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	4	document	document	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	5	dissemination	dissemination	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	8	WWW	WWW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	308	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	2	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	3	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	4	Bulletin	Bulletin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	6	1996	1996	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	309	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	2	7	7	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	4	Dario	Dario	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	5	Bonino	Bonino	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	7	Fulvio	Fulvio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	8	Corno	Corno	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	11	Giovanni	Giovanni	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	12	Squillero	Squillero	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	310	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	2	real	real	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	3	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	4	time	time	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	5	evolu-	evolu-	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	6	tionary	tionary	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	7	algorithm	algorithm	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	8	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	10	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	311	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	6	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	7	WIC	WIC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	8	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	9	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	11	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	12	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	14	Halifax	Halifax	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	16	Canada	Canada	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	18	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	312	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	2	8	8	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	4	Xin	Xin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	5	Chen	Chen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	7	Xiaodong	Xiaodong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	8	Zhang	Zhang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	313	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	1	Popularity	popularity	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	3	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	4	PPM	PPM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	5	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	6	An	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	7	effective	effective	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	9	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	10	technique	technique	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	11	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	12	high	high	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	13	accuracy	accuracy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	15	low	low	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	16	storage	storage	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	314	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	5	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	6	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	8	Parallel	Parallel	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	9	Processing	Processing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	11	Vancouver	Vancouver	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	13	Canada	Canada	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	15	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	315	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	2	9	9	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	4	Xin	Xin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	5	Chen	Chen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	7	Xiaodong	Xiaodong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	8	Zhang	Zhang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	316	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	1	Coordinated	coordinated	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	2	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	3	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	4	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	6	contents	content	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	317	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	1	Computer	Computer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	2	Communications	Communications	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	4	28:1947–1958	28:1947–1958	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	6	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	318	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	2	10	10	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	4	K.	K.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	5	Chinen	Chinen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	7	S.	S.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	8	Yamaguchi	Yamaguchi	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	319	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	1	An	an	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	2	interactive	interactive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	3	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	4	proxy	proxy	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	5	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	7	improvement	improvement	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	9	www	www	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	10	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	320	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	1	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	3	Seventh	Seventh	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	4	Annual	Annual	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	5	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	8	Internet	Internet	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	9	Society	Society	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	10	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	11	INET’97	INET’97	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	12	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	14	06/1997	06/1997	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	15	1997	1997	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	321	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	2	11	11	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	4	Wei	Wei	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	5	Chu	Chu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	7	Seung	Seung	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	9	Taek	Taek	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	10	Park	Park	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	322	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	1	Personalized	personalize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	2	recommendation	recommendation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	3	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	4	dynamic	dynamic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	5	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	6	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	7	predictive	predictive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	8	bilinear	bilinear	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	9	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	323	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	5	18th	18th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	6	Inter-	Inter-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	7	national	national	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	8	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	10	World	World	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	11	Wide	Wide	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	12	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	14	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	15	691–700	691–700	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	17	Madrid	Madrid	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	19	Spain	Spain	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	21	2009	2009	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	324	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	1	13	13	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	3	12	12	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	5	Edith	Edith	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	6	Cohen	Cohen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	8	Haim	Haim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	9	Kaplan	Kaplan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	325	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	1	Prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	3	means	mean	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	4	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	5	document	document	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	6	transfer	transfer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	7	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	9	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	10	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	11	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	12	reducing	reduce	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	13	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	14	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	326	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	1	Computer	Computer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	2	Networks	Networks	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	4	39(4):437–455	39(4):437–455	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	6	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	327	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	2	13	13	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	4	Edith	Edith	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	5	Cohen	Cohen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	7	Balachander	Balachander	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	8	Krishnamurthy	Krishnamurthy	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	11	Jennifer	Jennifer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	12	Rexford	Rexford	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	328	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	1	Efficient	efficient	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	2	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	3	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	4	predicting	predict	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	5	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	6	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	8	servers	server	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	329	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	6	INFOCOM	INFOCOM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	7	’	’	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	8	99	99	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	9	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	11	New	New	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	12	York	York	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	14	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	16	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	330	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	2	14	14	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	4	Brian	Brian	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	5	D.	D.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	6	Davison	Davison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	331	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	1	Predicting	predict	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	2	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	3	actions	action	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	4	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	5	HTML	html	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	6	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	332	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	2	Proceed-	Proceed-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	3	ings	ing	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	6	13th	13th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	7	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	8	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	10	Hypertext	Hypertext	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	12	Hypermedia	Hypermedia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	14	College	College	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	15	Park	Park	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	17	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	19	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	333	20	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	2	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	4	Brian	Brian	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	5	D.	D.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	6	Davison	Davison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	334	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	335	1	Learning	learning	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	335	2	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	335	3	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	335	4	patterns	pattern	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	335	5	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	2	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	3	Dynamics	Dynamics	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	4	-	-	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	5	Adapting	adapting	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	6	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	7	Change	Change	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	9	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	11	Size	Size	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	13	Topology	Topology	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	15	Use	Use	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	17	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	18	435–460	435–460	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	336	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	337	1	Springer	Springer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	337	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	337	3	2004	2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	337	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	2	16	16	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	4	Bernardo	Bernardo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	5	de	de	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	6	la	la	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	7	Ossa	Ossa	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	9	Ana	Ana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	10	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	12	Julio	Julio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	13	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	15	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	338	16	J.	J.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	339	1	A.	A.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	339	2	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	339	3	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	1	Referrer	referrer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	2	graph	graph	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	3	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	4	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	5	low	low	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	6	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	7	cost	cost	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	9	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	10	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	340	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	341	1	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	341	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	341	3	2010	2010	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	341	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	2	17	17	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	4	Josep	Josep	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	5	Domenech	Domenech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	7	Jose	Jose	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	8	A.	A.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	9	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	11	Julio	Julio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	12	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	15	Ana	Ana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	16	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	342	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	1	Speculative	speculative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	2	validation	validation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	4	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	5	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	7	further	further	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	8	reducing	reduce	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	10	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	11	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	12	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	13	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	343	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	5	9th	9th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	6	International	international	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	7	IFIP	IFIP	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	8	TC	TC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	9	6	6	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	10	Networking	Networking	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	11	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	13	Chennai	Chennai	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	15	India	India	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	17	2010	2010	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	344	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	2	18	18	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	4	Josep	Josep	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	5	Domenech	Domenech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	7	Jose	Jose	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	8	A.	A.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	9	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	11	Julio	Julio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	12	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	15	Ana	Ana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	16	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	345	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	1	Using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	2	current	current	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	3	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	4	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	5	structure	structure	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	6	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	7	improve	improve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	8	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	9	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	346	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	1	Computer	Computer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	2	Networks	Networks	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	4	54(9):1404	54(9):1404	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	5	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	6	1417	1417	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	8	2010	2010	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	347	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	2	19	19	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	4	Josep	Josep	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	5	Domènech	Domènech	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	7	Julio	Julio	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	8	Sahuquillo	Sahuquillo	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	10	José	José	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	11	A.	A.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	12	Gil	Gil	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	15	Ana	Ana	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	16	Pont	Pont	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	348	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	2	impact	impact	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	6	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	7	architecture	architecture	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	10	limits	limit	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	11	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	12	reducing	reduce	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	13	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	14	’s	’s	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	15	perceived	perceive	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	16	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	349	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	5	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	6	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	7	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	8	WIC	WIC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	9	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	10	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	11	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	12	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	13	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	14	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	15	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	17	Hong	Hong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	18	Kong	Kong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	20	China	China	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	22	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	350	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	2	20	20	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	4	Xing	Xing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	5	Dongshan	Dongshan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	7	Shen	Shen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	8	Junyi	Junyi	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	351	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	2	new	new	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	3	Markov	Markov	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	4	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	5	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	7	access	access	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	8	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	352	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	1	Computing	compute	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	2	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	3	Science	Science	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	5	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	7	4(6):34–39	4(6):34–39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	9	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	353	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	2	21	21	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	4	Dan	Dan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	5	Duchamp	Duchamp	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	354	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	355	1	Prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	355	2	hyperlinks	hyperlink	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	355	3	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	5	2nd	2nd	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	6	USENIX	USENIX	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	7	Symposium	Symposium	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	9	Internet	Internet	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	10	Technologies	Technologies	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	12	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	14	Boulder	Boulder	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	16	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	18	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	356	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	2	22	22	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	4	Li	Li	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	5	Fan	Fan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	7	Pei	Pei	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	8	Cao	Cao	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	10	Wei	Wei	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	11	Lin	Lin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	13	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	14	Quinn	Quinn	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	15	Jacobson	Jacobson	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	357	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	2	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	3	between	between	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	4	low	low	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	5	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	6	bandwidth	bandwidth	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	7	clients	client	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	9	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	10	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	11	Potential	potential	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	13	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	358	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	2	Proceed-	Proceed-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	3	ings	ing	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	5	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	6	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	7	SIGMETRICS	SIGMETRICS	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	8	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	9	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	10	Measurement	Measurement	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	12	Modeling	Modeling	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	13	Of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	14	Computer	Computer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	15	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	17	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	18	178–187	178–187	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	20	Atlanta	Atlanta	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	22	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	24	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	359	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	2	23	23	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	4	A	A	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	5	Georgakis	Georgakis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	7	H	H	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	8	Li	Li	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	360	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	1	User	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	2	behavior	behavior	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	3	modeling	modeling	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	5	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	6	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	7	specu-	specu-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	8	lative	lative	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	10	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	11	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	361	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	1	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	2	&	&	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	3	Knowledge	Knowledge	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	4	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	6	59:770	59:770	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	7	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	8	788	788	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	10	12/2006	12/2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	11	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	362	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	2	24	24	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	4	Sule	Sule	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	5	Gunduz	Gunduz	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	7	M.	M.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	8	Tamer	Tamer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	9	Zsu	Zsu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	363	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	2	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	3	page	page	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	4	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	5	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	6	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	8	click	click	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	9	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	10	stream	stream	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	11	tree	tree	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	12	representation	representation	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	13	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	14	user	user	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	15	behavior	behavior	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	364	16	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	2	KDD	KDD	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	3	’	'	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	4	03	03	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	5	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	6	Proceed-	Proceed-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	7	ings	ing	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	9	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	10	ninth	ninth	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	11	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	12	SIGKDD	SIGKDD	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	13	international	international	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	14	conference	conference	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	15	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	16	Knowledge	Knowledge	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	17	discovery	discovery	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	18	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	19	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	20	mining	mining	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	22	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	23	535–540	535–540	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	25	New	New	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	26	York	York	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	27	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	28	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	29	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	30	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	365	31	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	366	1	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	366	2	Press	Press	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	366	3	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	1	14	14	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	3	25	25	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	5	Yin	Yin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	6	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	7	Fu	Fu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	8	Huang	Huang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	10	Jhao	Jhao	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	11	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	12	Min	Min	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	13	Hsu	Hsu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	367	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	1	Mining	mining	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	2	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	3	logs	log	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	4	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	5	improve	improve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	6	hit	hit	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	7	ratios	ratio	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	8	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	9	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	11	caching	cache	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	368	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	5	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	6	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	7	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	8	WIC	WIC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	9	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	10	ACM	ACM	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	11	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	12	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	13	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	14	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	15	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	17	Compiegne	Compiegne	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	18	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	19	France	France	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	20	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	21	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	369	22	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	2	26	26	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	4	Yingyin	Yingyin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	5	Jiang	Jiang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	7	Min	Min	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	9	You	You	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	10	Wu	Wu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	13	Wei	Wei	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	14	Shu	Shu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	370	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	2	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	3	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	4	Costs	cost	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	6	bene-	bene-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	7	fits	fit	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	9	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	371	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	5	7th	7th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	7	Workshop	Workshop	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	9	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	10	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	11	Caching	Caching	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	12	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	13	Content	Content	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	14	Distribution	Distribution	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	16	Boulder	Boulder	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	18	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	20	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	372	21	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	2	27	27	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	4	Javed	Javed	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	5	I.	I.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	6	Khan	Khan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	8	Qingping	Qingping	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	9	Tao	Tao	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	373	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	1	Exploiting	exploit	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	2	webspace	webspace	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	3	organization	organization	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	4	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	5	ac-	ac-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	6	celerating	celerate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	8	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	374	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	6	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	7	WIC	WIC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	8	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	9	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	11	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	12	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	14	Halifax	Halifax	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	16	Canada	Canada	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	18	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	375	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	2	28	28	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	4	Yuna	Yuna	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	5	Kim	Kim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	7	Jong	Jong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	8	Kim	Kim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	376	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	2	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	3	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	4	display	display	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	5	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	6	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	7	predic-	predic-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	8	tion	tion	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	377	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	6	/	/	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	7	WIC	WIC	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	8	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	9	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	10	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	11	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	12	Intelligence	Intelligence	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	14	Halifax	Halifax	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	16	Canada	Canada	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	18	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	378	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	2	29	29	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	4	Thomas	Thomas	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	5	M.	M.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	6	Kroeger	Kroeger	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	8	Darrell	Darrell	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	379	9	D.E.	D.E.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	1	Long	long	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	3	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	4	Jeffrey	Jeffrey	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	5	C.	C.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	6	Mogul	Mogul	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	380	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	1	Exploring	explore	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	3	bounds	bound	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	6	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	7	reduction	reduction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	8	from	from	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	9	caching	cache	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	11	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	381	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	5	1st	1st	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	6	USENIX	USENIX	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	7	Symposium	Symposium	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	9	Internet	Internet	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	10	Technologies	Technologies	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	12	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	14	Monterey	Monterey	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	16	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	18	1997	1997	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	382	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	2	30	30	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	4	Kelvin	Kelvin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	5	Lau	Lau	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	7	Yiu	Yiu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	9	Kai	Kai	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	10	Ng	Ng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	383	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	2	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	3	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	4	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	6	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	7	manage-	manage-	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	8	ment	ment	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	9	system	system	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	10	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	11	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	12	detection	detection	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	13	theory	theory	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	384	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	5	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	6	Con-	Con-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	7	tent	tent	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	8	Caching	Caching	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	9	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	10	Distribution	Distribution	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	11	:	:	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	12	9th	9th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	13	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	14	Workshop	Workshop	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	15	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	16	WCW	WCW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	17	2004	2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	18	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	20	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	21	129–143	129–143	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	22	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	23	Beijing	Beijing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	24	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	25	China	China	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	26	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	27	2004	2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	385	28	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	2	31	31	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	4	Heung	Heung	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	5	Ki	Ki	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	6	Lee	Lee	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	8	Gopinath	Gopinath	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	9	Vageesan	Vageesan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	11	Ki	Ki	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	12	Hwan	Hwan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	13	Yum	Yum	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	14	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	15	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	16	Eun	Eun	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	17	Jung	Jung	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	18	Kim	Kim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	386	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	2	proactive	proactive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	3	request	request	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	4	distribution	distribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	5	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	6	prord	prord	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	7	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	8	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	10	log	log	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	11	mining	mining	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	12	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	13	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	14	cluster-	cluster-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	15	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	16	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	17	server	server	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	387	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	5	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	6	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	8	Parallel	Parallel	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	9	Processing	Processing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	10	(	(	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	11	ICPP’06	ICPP’06	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	12	)	)	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	14	Columbus	Columbus	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	16	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	18	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	388	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	2	32	32	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	4	Nizar	Nizar	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	5	R.	R.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	6	Mabroukeh	Mabroukeh	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	8	Christie	Christie	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	9	I.	I.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	10	Ezeife	Ezeife	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	389	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	1	Semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	3	rich	rich	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	4	markov	markov	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	5	models	model	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	6	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	8	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	390	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	5	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	7	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	9	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	10	Mining	Mining	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	11	Workshops	Workshops	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	13	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	14	465	465	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	15	–	–	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	16	470	470	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	17	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	18	Miami	Miami	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	20	FL	FL	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	21	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	22	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	23	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	24	2009	2009	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	391	25	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	392	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	392	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	392	3	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	392	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	2	33	33	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	4	Evangelos	Evangelos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	5	Markatos	Markatos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	7	Catherine	Catherine	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	8	Chronaki	Chronaki	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	393	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	2	top-10	top-10	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	3	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	4	to	to	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	5	pre-	pre-	RB
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	6	fetching	fetch	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	8	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	394	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	5	INET	INET	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	6	’	'	''
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	7	98	98	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	9	Geneva	Geneva	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	11	Switzerland	Switzerland	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	13	1998	1998	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	395	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	2	34	34	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	4	Bamshad	Bamshad	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	5	Mobasher	Mobasher	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	396	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	2	Adaptive	Adaptive	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	3	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	5	volume	volume	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	6	LNCS	LNCS	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	7	4321	4321	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	9	chapter	chapter	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	10	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	11	Mining	Mining	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	12	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	13	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	14	Personalization	Personalization	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	16	pages	page	VBZ
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	17	90–135	90–135	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	397	18	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	398	1	Springer	Springer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	398	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	398	3	2007	2007	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	398	4	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	2	35	35	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	4	Alexandros	Alexandros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	5	Nanopoulos	Nanopoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	7	Dimitrios	Dimitrios	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	8	Katsaros	Katsaros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	11	Yannis	Yannis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	12	Manolopoulos	Manolopoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	399	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	2	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	3	mining	mining	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	4	algorithm	algorithm	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	5	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	6	generalized	generalize	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	8	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	400	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	401	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	401	2	Trans	Trans	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	401	3	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	402	1	Knowl	Knowl	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	402	2	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	1	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	2	Eng	Eng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	3	.	.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	5	15(5):1155–1169	15(5):1155–1169	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	7	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	403	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	2	36	36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	4	Venkata	Venkata	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	5	N.	N.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	6	Padmanabhan	Padmanabhan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	8	Jeffrey	Jeffrey	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	9	C.	C.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	10	Mogul	Mogul	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	404	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	1	Using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	2	predictive	predictive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	3	pre-	pre-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	4	fetching	fetch	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	5	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	6	improve	improve	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	7	World	World	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	8	Wide	Wide	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	9	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	10	latency	latency	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	405	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	1	Computer	computer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	2	Communication	Communication	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	3	Review	Review	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	5	26(3):22–36	26(3):22–36	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	7	1996	1996	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	406	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	2	37	37	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	4	G	G	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	5	Pallis	Pallis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	7	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	8	Vakali	Vakali	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	11	J	J	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	12	Pokorny	Pokorny	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	407	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	2	clustering	clustering	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	3	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	4	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	5	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	6	scheme	scheme	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	7	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	8	a	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	10	cache	cache	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	11	environment	environment	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	408	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	1	Computers	Computers	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	2	&	&	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	3	Electrical	Electrical	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	4	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	6	34:309	34:309	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	7	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	8	323	323	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	10	07/2008	07/2008	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	11	2008	2008	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	409	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	1	15	15	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	2	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	3	38	38	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	4	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	5	Themistoklis	Themistoklis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	6	Palpanas	Palpanas	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	7	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	8	Alberto	Alberto	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	9	Mendelzon	Mendelzon	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	410	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	1	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	2	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	3	using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	4	par-	par-	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	5	tial	tial	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	6	match	match	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	7	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	411	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	2	Proceedings	proceeding	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	5	4th	4th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	7	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	8	Caching	Caching	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	9	Workshop	Workshop	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	11	San	San	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	12	Diego	Diego	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	14	USA	USA	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	16	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	412	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	2	39	39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	4	A	A	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	5	Pons	Pons	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	413	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	1	Improving	improve	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	2	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	3	performance	performance	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	5	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	7	object	object	VBP
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	8	retrieval	retrieval	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	414	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	1	Journal	Journal	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	3	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	5	Software	Software	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	7	74:303	74:303	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	8	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	9	311	311	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	11	02/2005	02/2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	12	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	415	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	2	40	40	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	4	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	5	Pons	Pons	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	416	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	1	Semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	2	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	3	objects	object	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	4	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	5	slower	slow	JJR
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	7	site	site	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	8	pages	page	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	417	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	1	Journal	Journal	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	3	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	5	Software	Software	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	7	79:1715	79:1715	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	8	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	9	1724	1724	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	11	12/2006	12/2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	12	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	418	13	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	2	41	41	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	4	Michael	Michael	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	5	Rabinovich	Rabinovich	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	7	Oliver	Oliver	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	8	Spatscheck	Spatscheck	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	419	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	420	1	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	420	2	Caching	Caching	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	420	3	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	420	4	Replication	Replication	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	420	5	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	1	Addison	Addison	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	3	Wesley	Wesley	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	5	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	421	6	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	422	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	422	2	42	42	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	422	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	422	4	S.K.	S.K.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	423	1	Rangarajan	Rangarajan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	423	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	423	3	V.V.	V.V.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	424	1	Phoha	Phoha	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	424	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	424	3	K.S.	K.S.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	425	1	Balagani	Balagani	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	425	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	425	3	R.R.	R.R.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	1	Selmic	selmic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	3	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	4	S.S.	S.S.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	5	Iyen-	Iyen-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	6	gar	gar	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	426	7	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	1	Adaptive	adaptive	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	2	neural	neural	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	3	network	network	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	4	clustering	clustering	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	5	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	6	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	7	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	427	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	1	Computer	computer	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	2	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	3	37:34	37:34	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	4	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	5	40	40	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	7	04/2004	04/2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	8	2004	2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	428	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	2	43	43	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	4	Stuart	Stuart	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	5	Schechter	Schechter	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	7	Murali	Murali	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	8	Krishnan	Krishnan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	11	Michael	Michael	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	12	D.	D.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	13	Smith	Smith	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	429	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	1	Using	use	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	2	path	path	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	3	profiles	profile	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	4	to	to	TO
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	5	predict	predict	VB
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	6	http	http	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	7	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	430	8	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	5	7th	7th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	7	World	World	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	8	Wide	Wide	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	9	Web	Web	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	10	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	12	Brisbane	Brisbane	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	14	Australia	Australia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	16	1998	1998	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	431	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	2	44	44	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	4	Antonis	Antonis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	5	Sidiropoulos	Sidiropoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	7	George	George	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	8	Pallis	Pallis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	10	Dimitrios	Dimitrios	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	11	Katsaros	Katsaros	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	13	Konstantinos	Konstantinos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	14	Sta-	Sta-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	15	mos	mos	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	16	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	17	Athena	Athena	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	18	Vakali	Vakali	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	19	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	20	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	21	Yannis	Yannis	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	22	Manolopoulos	Manolopoulos	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	432	23	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	1	Prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	2	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	3	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	4	distribution	distribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	5	networks	network	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	6	via	via	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	7	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	8	communities	community	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	9	identification	identification	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	10	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	11	outsourcing	outsourcing	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	433	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	1	World	World	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	2	Wide	wide	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	3	Web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	4	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	5	11:39	11:39	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	6	–	–	SYM
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	7	70	70	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	9	3/2008	3/2008	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	10	2008	2008	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	434	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	2	45	45	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	4	Wei	Wei	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	5	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	6	Guang	Guang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	7	Teng	Teng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	9	Cheng	Cheng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	10	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	11	Yue	Yue	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	12	Chang	Chang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	14	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	15	Ming	Ming	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	16	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	17	Syan	Syan	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	18	Chen	Chen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	435	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	1	Integrating	integrate	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	2	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	3	caching	caching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	5	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	6	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	7	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	8	client	client	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	9	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	10	side	side	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	11	proxies	proxy	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	436	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	2	Transactions	transaction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	3	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	4	Parallel	Parallel	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	6	Distributed	Distributed	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	7	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	9	16(5):444–455	16(5):444–455	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	11	2005	2005	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	437	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	2	46	46	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	4	Arun	Arun	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	5	Venkataramani	Venkataramani	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	7	Praveen	Praveen	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	8	Yalagandula	Yalagandula	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	9	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	10	Ravindranath	Ravindranath	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	11	Kokku	Kokku	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	13	Sadia	Sadia	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	14	Sharif	Sharif	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	16	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	17	Mike	Mike	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	18	Dahlin	Dahlin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	438	19	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	1	The	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	2	potential	potential	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	3	costs	cost	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	4	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	5	benefits	benefit	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	6	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	7	long	long	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	8	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	9	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	10	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	11	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	12	content	content	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	13	distribution	distribution	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	439	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	1	Computer	Computer	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	2	Communications	Communications	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	3	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	4	25:367	25:367	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	5	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	6	375	375	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	8	2002	2002	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	440	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	2	47	47	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	4	Bin	Bin	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	5	Wu	Wu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	6	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	7	Ajay	Ajay	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	8	D.	D.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	9	Kshemkalyani	Kshemkalyani	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	441	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	1	Objective	objective	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	2	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	3	optimal	optimal	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	4	algorithms	algorithm	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	5	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	6	long-	long-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	7	term	term	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	8	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	9	prefetching	prefetche	VBG
work_jsnmrvhjlfgxfkyxshfjgxeayu	442	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	2	Transactions	transaction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	3	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	4	Computers	Computers	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	5	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	6	55(1):2–17	55(1):2–17	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	7	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	8	2006	2006	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	443	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	2	48	48	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	4	Cheng	Cheng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	5	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	6	Zhong	Zhong	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	7	Xu	Xu	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	8	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	444	9	T.I.	T.I.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	445	1	Ibrahim	Ibrahim	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	445	2	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	2	keyword	keyword	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	3	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	4	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	5	semantic	semantic	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	6	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	7	approach	approach	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	8	in	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	9	internet	internet	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	10	news	news	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	11	services	service	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	446	12	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	1	IEEE	IEEE	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	2	Transactions	transaction	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	3	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	4	Knowledge	Knowledge	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	5	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	6	Data	Data	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	7	Engineering	Engineering	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	9	16:601	16:601	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	10	–	–	:
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	11	611	611	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	12	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	13	05/2004	05/2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	14	2004	2004	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	447	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	2	49	49	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	4	Qiang	Qiang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	5	Yang	Yang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	7	Joshua	Joshua	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	8	Zhexue	Zhexue	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	9	Huang	Huang	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	12	Michael	Michael	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	13	Ng	Ng	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	448	14	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	1	A	a	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	2	data	datum	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	3	cube	cube	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	4	model	model	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	5	for	for	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	6	prediction	prediction	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	7	-	-	HYPH
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	8	based	base	VBN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	9	web	web	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	10	prefetching	prefetching	NN
work_jsnmrvhjlfgxfkyxshfjgxeayu	449	11	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	1	Journal	Journal	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	2	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	3	Intelligent	Intelligent	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	4	Information	Information	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	5	Systems	Systems	NNPS
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	7	20(1):11–30	20(1):11–30	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	8	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	9	2003	2003	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	450	10	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	1	[	[	-LRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	2	50	50	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	3	]	]	-RRB-
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	4	Ingrid	Ingrid	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	5	Zukerman	Zukerman	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	6	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	7	David	David	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	8	W.	W.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	9	Albrecht	Albrecht	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	10	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	11	and	and	CC
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	12	Ann	Ann	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	13	E.	E.	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	14	Nicholson	Nicholson	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	451	15	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	1	Predict-	Predict-	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	2	ing	ing	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	3	users	user	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	4	’	’	POS
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	5	requests	request	NNS
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	6	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	7	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	8	WWW	WWW	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	452	9	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	1	In	in	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	2	Proceedings	Proceedings	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	3	of	of	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	4	the	the	DT
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	5	7th	7th	JJ
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	6	International	International	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	7	Conference	Conference	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	8	on	on	IN
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	9	User	User	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	10	Modeling	Modeling	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	11	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	12	Banff	Banff	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	13	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	14	Canada	Canada	NNP
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	15	,	,	,
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	16	1999	1999	CD
work_jsnmrvhjlfgxfkyxshfjgxeayu	453	17	.	.	.
work_jsnmrvhjlfgxfkyxshfjgxeayu	454	1	16	16	CD