Matchmaking of Semantic Web Services Using Semantic-Distance Information

Abstract

This paper mainly focuses on proposing efficient and extensible matchmaking architecture. We ran an improved algorithm using the semantic-distance information, based on OWL-S, UDDI and semantic web architectures. Current matchmaking architectures and algorithms lack vision and they are unable to use all available information. However, our proposed architecture uses information such as path-length of the ontological tree nodes and partial results sets for composing required service even no exact match is found. Semantic-distance information may be used as selection criteria and it provides accuracy in service selection. To define concept-similarity rating by ontology managers or local users may provide a way for service selection. We can then gather second level of information other than the pre-defined match levels such as exact, subsume or plug-in. We exploit this fact in our architecture and algorithm by providing a layered and extended architecture. Different filtering layers and different specifications are applied in the matchmaking to give an extendable architecture.