id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_qbzexnx22zfibibhrvs5slvbk4	Susan Gauch	An expert system for automatic query reformation	1993	30	.pdf	application/pdf	10921	924	66	prototype automatically reformulates queries to improve the search results, and the number of queries necessary to find relevant passages compared with the user In our system, queries are reformulated by a knowledge-based online search retrieved passages are presented to the user in rank-order. for the expert system's query reformulation knowledge base were: 1) each search To broaden a query, search terms are added to the positive concepts, whereas narrowing a query adds search terms to negative concepts. more likely to retrieve relevant passages than broadening with the parent terms, original query would retrieve only one passage, so the expert system would attempt user's original search terms; the number of occurrences for each word type from the Hypothesis 3: The expert system can rank the passages retrieved by the search in function, and did no query reformulation or ranking of retrieved passages. the number of user indicated relevant passages in the final retrieval set for the	./cache/work_qbzexnx22zfibibhrvs5slvbk4.pdf	./txt/work_qbzexnx22zfibibhrvs5slvbk4.txt