Multilingual enrichment of an ontology of cardio-vascular diseases

Concept-based information retrieval offers a higher abstraction level than the classical keyword-based approach. A multilingual terminology makes possible language-independent indexing and querying. The conceptual structure with its associated terminology is called an ontology. This paper describes a methodology to enrich an existing ontology represented in OWL with terms in different languages, and its application to a MeSH-based ontology of the cardio-vascular domain. The first enrichment phase is automatic and makes use of the UMLS multilingual lexicon. The second phase is partially manual and relies on a corpus of texts. In the Noesis European project, where this work was carried on, the multilingual ontology was used for automatic indexing and manual semantic annotation of cardio-vascular documents.