On Aligning Interesting Parts of Ontologies

Abstract

Ontology merging/alignment is one of the most important tasks in ontology engineering. It is imposed by the decentralized nature of both the WWW and the Semantic Web, where heterogeneous and incompatible ontologies can be developed by different communities. Usually, ontology merging/alignment is based on an ontology mapping that has been established in a previous phase. In this paper, we define a new problem within the alignment process: the problem of detecting and then updating only interesting parts of an ontology, based on the knowledge included in another one. To this end, we define and evaluate a number of different measures of interestingness of parts of ontologies. We also present experimental results for their evaluation on test ontologies.