Enhancing Query Reformulation Performance by Combining Content and Hypertext Analyses

Information retrieval techniques play a critical role in the development of the information systems. Different searches have focused on the way of improving the retrieval effectiveness. Query expansion via relevance feedback is a good technique that proved to be a good way to improve the retrieval performance. In this paper, we investigate new methods to improve the query reformulation process. A two step process is employed to reformulate query. In a preliminary step, a local set of documents is built from the retrieved result. In a second step, a co-occurrence analysis is performed on the local document set to deduce the terms to be used for the query expansion. To build the local set we use firstly a content-based analysis. It is a similarity study between the retrieved documents and the query. The second method combines content and hypertext analyses to achieve the local set construction. The TREC frame is used to evaluate the proposed processes.