A Hybrid Ontology Mediation Approach for the Semantic Web

This article introduces a hybrid ontology mediation approach for the Semantic Web. It combines both syntactic and semantic matching measures to provide better results for matching data labels. Although ontologies are meant to provide a shared conceptualization of the world, the development practices, lack of standards, and subjective naming conventions today, create data heterogeneity problems among ontologies. This is a significant problem particularly for ontologies of similar domains. Mediation techniques at present focus mainly on syntactic matching and our premise is that a hybrid approach would be a better solution to this problem. We provide empirical evidence with hypothesis tests and also provide several new measures such as relevance, reliability, and precision to validate our approach. We also introduce a detailed mapping algorithm.