An Information Retrieval Expansion Model Based on Quasi-Clique

Query expansion is an important technology for improving retrieval performance in information retrieval. Many Studies have found contexts within query that strongly influence the interpretation of a query. In this paper, we propose the graph mining technique called Quasi-Clique as query context in Markov network retrieval model. Our approach exploits contextual information mined from the term Markov network for per query term in addition to syntactic similarity. The proposal in our work is benefit to select more relevant terms for query expansion and to improve precision in information retrieval. Keywords—information retrieval; query expansion; quasiclique; Markov network retrieval