Similarity computation by ontology merging system: DKP-OM

One of the core tasks of mapping and merging of multiple ontologies is to produce accurate, consistent and coherent merged global ontology that promotes interoperability among heterogeneous multi-vendors semantic-based systems. Current systems for ontology mapping and merging are very restricted in terms of resolving mismatches or proposing accurate matches with no or minimum human intervention. The suggestions made by these systems do not consider all information available in the semantic knowledge of the ontologies. We developed a semantic based ontology merging system, DKP-OM that employs almost all the semantics provided in the ontologies. This paper discusses the similarity computation factors considered by DKP-OM for providing complete, consistent and coherent merged ontology. These factors find the correspondences between the concepts of ontologies and check the consistency of initial mappings found. The concept of validation during the initial stages of ontology mapping not only distinguishes our system from the existing ones´ but also reduces the users´ dependability for validating the consistency of the generated mappings.