ONTOLOGIES ALIGNMENT TOWARDS A HYBRID APPROACH (A STATE OF THE ART)

Abstract

Ontologies are concepts that can represent forms of organization and intelligent control of knowledge. The diversity of ontologies poses a problem of heterogeneity, hence the need for the development of techniques and Tools aimed at promoting a certain interoperability of these ontologies and their union. Alignment between two ontologies (matching or mapping) consists of producing a set of matches between entities. These entities can be concepts, properties, or instances. As the evolution of these ontologies must be ensured, highly elaborate artificial intelligence techniques make it possible to ensure machine learning of the alignment process. This article presents a state-of-the-art of ontology alignment combined with artificial intelligence techniques that irreversibly improve the alignment process. The results known to date show several of several artificial intelligence tools implemented as part of the alignment and which have produced convincing results. In view of the limitations presented by the proposed approaches, we have opted to use genetic algorithms inspired by natural selection to optimize the results of the OWL ontology alignments.