A Framework for Improving the Performance of Ontology Matching Techniques in Semantic Web

Abstract

Ontology matching is the process of finding correspondences between semantically related entities of different ontologies. We need to apply this process to solve the heterogeneity problems between different ontologies. Some ontologies may contain thousands of entities which make the ontology matching process very complex in terms of space and time requirements. This paper presents a framework that reduces the search space by removing entities (classes, properties) that have less probability of being matched. In order to achieve this goal we have introduced a matching strategy that uses multi matching techniques specifically; string, structure, and linguistic matching techniques. The results obtained from this framework have indicated a good quality matching outcomes in a low time requirement and a low search space in comparisons with other matching frameworks. It saves from the search space from (43% - 53%), and saves on the time requirement from (38% - 45%).