Clustering Relations WITHIN OWL Ontologies

C.R. Kothari and D.J. Russomanno (USA)


Semantic Web, Ontologies, Relations, Clusters


This paper discusses a prototype that uses the explicitly asserted intrinsic semantics of relations in OWL ontologies to group them into clusters based upon quantitative measures of their mutual similarities. The intrinsic nature of relations can be expressed using a collection of relation elements. Relation elements can be used to quantitatively analyze the extent of the similarities and dissimilarities among a set of relations. With quantitative measures of the similarity between every pair of relations from an ontology, methods can be used to partition the relation space of the ontology into clusters. Relation clusters, both within and among ontologies, may help in knowledge reuse and interoperability in a distributed environment such as the Semantic Web. A prototype has been developed to use the assertions of the intrinsic nature of relations captured in OWL Full ontologies to create relation clusters. Results from testing this prototype on a high-level ontology are presented and analyzed in this paper.

Important Links:

Go Back