A Term-Reweighting Method for Query Expansion
Chang Wang, Peiying Zhang, Baochuan Han · 2010
In this paper, we propose a term-reweighting method for query expansion. With the help of this approach we can extract high-quality query expansion terms from the entire document set. In this approach, the users' initial queries and the documents during the user relevance feedback are represented as vectors respectively, and the similarity between them can be obtained by calculating the vector similarity. Then we reweight the terms through one single document and the entire document set respectively, the final weight of the term can be obtained by the combination of the above weights, finally the terms with larger weight will be selected as query expansion terms. Experimental results show that the proposed method is quite reasonable and can improve the precision and recall significantly.