Thesaurus Based on Grey Theory for Interactive Query Expansion

With the increasing availability of information on the WWW (World Wide Web), it becomes more important and feasible to retrieve information efficiently. Query expansion is the process of supplementing an original query with additional terms in order to refine a search and increase retrieval effectiveness. If the query expansion is interactive, then the user and the system work together to expand the query. A well-designed thesaurus not only can represent the semantic knowledge among terms, but also can help to improve the retrieval performance in search engine. In previous researches, many thesauruses are represented as symmetric type, and this lead to some unreasonable results. Therefore, in this study, we propose an interactive searching scheme that aims to provide users an easy way to retrieve their desired information. A non-symmetric association thesaurus, which is generated by Sparse Grey Relational Analysis (SGRA) algorithm we proposed, is involved in query expansion process. Also, the SGRA algorithm improves the problem that traditional grey relational analysis cannot deal with sparse matrix, and to reduce the noise term. Experiment show that the searching scheme of expanding query by association thesaurus has improved retrieval performance slightly. We also find that the correctness of word segmentation and the completeness of dataset have a great impact on the quality of association thesaurus.