Combining knowledge-based methods to refine and expand queries in medicine

Abstract

Information retrieval remains problematic in spite of the numerous existing search engines. It is the same problem for health information retrieval. We propose in this paper to combine three knowledge-based methods to enhance information retrieval using query expansion in the context of the CISMeFproject (Catalogue and Index of French-speaking Medical Sites) in which the resources are indexed according to a structured terminology of the medical domainand a set of metadata. The first method consists of building and using morphological knowledge of the terms. The second method consists of extractingassociation rules between terms by applying a data mining technique over the indexed resources. The last method consists of formalizing the terminologyusing the OWL-DL language to benefit from its powerful reasoning mechanisms. We describe how these methods could be used conjointly in theKnowQuE prototype (Knowledge-based Query Expansion) and we give some preliminary results.