Conceptual Structure Matching using a Bayesian Framework in a Conceptual Indexing. Application to Medical Domain with Multilingual Documents and UMLS Meta-thesaurus
Abstract
Information Retrieval Systems that compute a matching between a document and a query based on words intersection, cannot reach relevant documents that do not share any terms with the query. The objective of this master thesis is to propose a solution to this problem in the context of conceptual indexing. We study an ontology based matching that exploit links between concepts. We propose a model that exploits the weighted links of ontology. We also propose to extend the links of the ontology to reflect the structural ambiguity of some concepts. A validation of our proposal is made on the test collection ImagCLEFMed 2005 and the external resource UMLS 2005.