A Robust Approach to the Ontology Matching Problem
Allaoua Refoufi, Achref Benarab · 2018
Ontology matching is the process that identifies correspondences between similar concepts in two different ontologies of the same domain of discourse to solve knowledge heterogeneous problems. We propose an automatic similarity based matching algorithm that exploits almost all types of entities descriptions as well as their relations to effectively compute the correspondences between the two to be matched ontologies. The iterative algorithm computes each measure of similarity separately and then aggregates them in a linear combination to compose the final similarity score. The measures used deal with linguistic, semantic, and structural as well as many other measures to gain efficiency. We also include a new similarity measure based on dynamic programming in conjunction with known measures to refine the similarity process. Finally, we provide comparative experimental results in support of our method on several well-known ontology benchmarks recommended by the OAEI1. The results obtained are shown to be quite superior compared to the state-of-the-art ontology matching systems.