Paper: Learning Web Query Patterns for Imitating Wikipedia Articles

ACL ID C10-2141
Title Learning Web Query Patterns for Imitating Wikipedia Articles
Venue International Conference on Computational Linguistics
Session Poster Session
Year 2010
Authors

This paper presents a novel method for ac- quiring a set of query patterns to retrieve documents containing important informa- tion about an entity. Given an existing Wikipedia category that contains the tar- get entity, we extract and select a small set of query patterns by presuming that formulating search queries with these pat- terns optimizes the overall precision and coverage of the returned Web informa- tion. We model this optimization prob- lem as a weighted maximum satisfiabil- ity (weighted Max-SAT) problem. The experimental results demonstrate that the proposedmethodoutperforms othermeth- ods based on statistical measures such as frequency and point-wise mutual informa- tion (PMI), which are widely used in rela- tion extraction.