Converting Classiflcations into OWL Ontologies

Fausto Giunchiglia,Ilya Zaihrayeu,Feroz Farazi

2008-01-01

Abstract:Classiflcation schemes, such as the DMoZ web directory, provide a convenient and intuitive way for humans to access classifled contents. While being easy to be dealt with for humans, classiflcation schemes remain hard to be reasoned about by automated software agents. Among other things, this hardness is conditioned by the ambiguous na- ture of the natural language used to describe classiflcation categories. In this paper we describe how classiflcation schemes can be converted into OWL ontologies, thus enabling reasoning on them by Semantic Web applications. The proposed solution is based on a two phase approach in which category names are flrst encoded in a concept language and then, together with the structure of the classiflcation scheme, are converted into an OWL ontology. We demonstrate the practical applicability of our approach by showing how the results of reasoning on these OWL ontologies can help improve the organization and use of web directories.