Abstract

In the field of information retrieval, the information of term frequency contained in relevance feedback has been widely used. However, the analysis and application of term frequency does not cover the semantic meaning of the terms, which could make the retrieval results deviate from the user’s searching goal. Consider the semantic meaning of the terms, Wille (1992) had proposed a structured view in the dealing with the term relationships of the terms in the retrieval documents. To enhance the effectiveness of information retrieval by the dealing with the mentioned information of term hierarchy relationship, this study has developed a method of query expansion to extract and apply this information contained in relevance feedback first, and then conducted some formal tests to verify the efficiency of the method in the re-ranking of the retrieved documents. The results of the formal tests show that the proposed method of query expansion is more effective than the Rocchio’s query expansion algorithm. The contribution of this study is the disclosure of the applicability of the information of term hierarchy relationship contained in relevance feedback, and the demonstration of the application of this information.

Share

COinS