Exploiting links to improve search in XML documents

Abstract

This paper describes an approach that ex-ploits links in XML retrieval. The proposed approach con-sists of reranking the set of documents returned for agiven query by considering three sources of evidencenamely, the relevance scores w.r.t query of a documentneighbours, the text of the anchor links, and the docu-ment title tag. Our approach is evaluated on INEX 2006collection. The results showed significant improvementsof the retrieval performances