Exploiting and Combining Multiple Resources for Query Expansion in Cross-Language Information Retrieval

概要

論文の詳細を見る
As Internet resources become accessible worldwide, the need to develop methods in Cross-Language Information Retrieval for different languages becomes increasingly important. In the present paper, we focus on query expansion techniques to improve the effectiveness of information retrieval. Combination of the dictionary-based translation and statistics-based disambiguation approaches is indispensable in overcoming query translation ambiguity. We therefore propose herein a model, which uses multiple sources for query reformulation, organization, translation and disambiguation, to select target translations and retrieve requested information. Relevance feedback or thesaurus-based expansion, as well as a new feedback strategy, which is based on the extraction of domain keywords to expand an original query, are introduced and evaluated. We tested the effectiveness of the proposed combined method using an application of French-English information retrieval. Experiments using the TREC data collection revealed the proposed combination of disambiguation and query expansion techniques to be highly effective.
社団法人情報処理学会の論文
2002-09-15