Learning concept hierarchies from textual resources for ontologies construction

摘要: Ontologies play a very important role in knowledge management and the Semantic Web, their use has been exploited many current applications. are especially useful because they support exchange sharing of information. Ontology learning from text is process deriving high-level concepts relations. An task ontology to obtain set representative model domain organize them into hierarchical structure (taxonomy) unstructured In building taxonomy, identification hypernym/hyponym relations between terms essential. How automatically build appropriate represent information contained texts challenging task. This paper presents novel method obtain, texts, taxonomic relationships specific domain. approach builds concept hierarchy specific-domain corpus by using clustering algorithm, linguistic patterns, additional contextual extracted Web that improves discovery most relationships. A experiments were carried out four different corpora. We evaluated quality constructed taxonomies against gold standard ontologies, show promising results.