Document expansion based on clique for Markov network information retrieval model

Document Expansion is an effective approach to solve the common problem of word mismatch in information retrieval. This paper proposes a novel document expansion method based on clique that gotten from the document space in Markov network built from corpus. As demonstrated by our experimental results on ADI, MED and CRAN data set, the method which integrates clique into document Expansion can achieve better query performance.