Computing similarity of semantic web services in semantic nets with multiple concept relations

The similarity of semantic web services is measured by matching service descriptions, which mostly depends on the understanding of their ontological concepts. Computing concept similarity on the basis of heterogeneous ontologies is still a problem. The current efforts only consider single hierarchical concept relations, which fail to express rich and implied information on concepts. Similarity under multiple types of concept relations as required by many application scenarios still needs to be investigated. To this end, first an original ontological concept similarity algorithm in a semantic net is proposed taking multiple concept relations into consideration, particularly fuzzy-weight relations between concepts. Then, this algorithm is employed to promote computing the similarity of semantic web services. An experimental prototype and detailed empirical discussions are presented, and the method is validated in the framework of web service selection.