Similarity Computation in Semantic Nets based on Multiple Concept Relations

Due to the scale and the decentralized nature of the Semantic Web, heterogeneities among concepts used are the rule rather than the exception. To provide a meaningful basis for tasks like reasoning over or discovery of semantic web services, similarity relationships must be defined among them in a semi-automatic way. Similarity computation schemes with single hierarchical concept relations are heavily studied, while the investigation of similarity under multiple concept relations as required in many application scenarios for semantic web services is still missing. Therefore, a novel algorithm was devised computing similarity in semantic nets by considering multiple concept relations, especially the fuzzy similarity relations between concepts. Empirical results obtained with this algorithm are presented.