Thanh-Le Bach and Rose Dieng-Kuntz
OWL becomes nowadays a more and more widely-used language for representing ontologies. The number of OWL ontologies increasing in direct ratio to the development of the Semantic Web leads to the heterogeneity problem. The same concepts may be modeled differently, using different terms and different positions in concept hierarchy. The task of identifying similar entities (concepts, relations or individuals) in different ontologies becomes then crucial for the success of information integration systems, instance transformation. In this paper, we propose a new similarity measure for comparing entities in different OWL DL ontologies. This measure is designed so as to enable extraction of information encoded in OWL entity descriptions and to take into account the underlying meaning of OWL primitives. We propose a variable weighting scheme for combining more efficiently component similarities calculated from components in entity descriptions.