Improving Automatic Query Expansion

Most casual users of IR systems type short queries. Recent research has shown that adding new words to these queries via adhoc feedback improves the retrieval e ectiveness of such queries. We investigate ways to improve this query expansion process by re ning the set of documents used in feedback. We start by using manually formulated Boolean lters along with proximity constraints. Our approach is similar to the one proposed by Hearst 12]. Next, we investigate a completely automatic method that makes use of term cooccurrence information to estimate word correlation. Experimental results show that re ning the set of documents used in query expansion often prevents the query drift caused by blind expansion and yields substantial improvements in retrieval e ectiveness, both in terms of average precision and precision in the top twenty documents. More importantly, the fully automatic approach developed in this study performs competitively with the best manual approach and requires little c...

• Type: Conference
• Year: 1998
• Where: SIGIR
• Authors: Mandar Mitra, Amit Singhal, Chris Buckley
• Keywords: Information Management, Most Casual Users, Query Expansion, Query Expansion Process, SIGIR 1998