Abstract:

Ontology clustering is an effective assistant means to integrate ontologies.However, studies on ontology clustering are rare.An ontology similarity computing method based on semantics is proposed.The method considers not only the concept, but also the attribute set and related concept set.Basic similarity threshold of concept is used to control probability to compute similarity of attribute sets.Semantic radius is used to control the scope of related concepts. Similarity matrix of ontology clustering is calculated using an ontology similarity method.An agglomerative hierarchical clustering algorithm is used for ontology clustering.Experiments show that semantic based ontology clustering can give satisfied results, and its performance is better than ontology clustering using the oMAP similarity method.The method has been applied in an aeronautical institute and played a special role in ontology integration.

Key words: Ontology clustering, semantics, similarity