Construction et enrichissement d'une ontologie à partir d'un corpus de textes

In this paper, we propose a construction and enrichment process of ontologies from texts. Ontologies are formal structures in which the concepts of a domain and relationships between them are formally defined. An increasing number of papers are dealing with ontologies but their construction and enrichment processes still a problem. Our construction approach is based on terms’s research into texts. It supposes that the frequent associations of two terms in some syntaxic structures could show a semantic relationship and then constitute the elements that could be integreted into an ontology. Terms and syntaxic structures identification is achieved through a partial and robust syntaxic analyser. These elements constitute the database on which operates the extraction process –extraction of frequent itemset– to extract regularities. MOTS-CLÉS : Ontologie, motifs fréquents, fouilles de textes