The Theme-Mine in Query Expansion

Query expansion (QE) is one of the most important mechanisms in the information retrieval field. A typical QE method usually redefines the query with related terms. These techniques suffer from retrieval performance degradation due to imprecise choice of query´s additive terms in the QE process. Semantic expansion is proposed to improve query precision, while the expansion concept is hard to be obtained due to less information. In this paper we introduce a novel theme mine technique which finding the expansion theme by users log analysis and topic clustering. The experimental results show that the theme, which mined by this method, exactly coincides with the user´s real intent, thus improve the power of query expansion.