Research On Application Of Ontology Mapping In Semantic Web Service Discovery

Semantic Web Services combines Semantic Web technology with Web Services technology,utilizes Ontology to describe Web services and Web resources.Service discovery is an important part of Semantic Web Services.Finding out a service that satisfies customers is the prerequisite of achieving service calling and service composition.There is ontology heterogeneity problem while achieving service matching in the phase of Semantic Web Service Discovery,and this problem could hamper the interoperability of Semantic Web Services,ontology mapping could solve it effectively.With a detailed survey about the necessary theoretic basis and the highly related fundamental technologies,the problem of ontology heterogeneity in Semantic Web Services was analyzed,various ontology mapping approaches and existing frameworks of service discovery were compared respectively,and Mediation technology proposed by WSMO organization was deeply studied,and then,a framework of Semantic Web Service Discovery based on Ontology Mapping was proposed,and the mechanism of its ontology mapping processor was elaborated.A prototype system OMSWSD(Ontology Mapping based Semantic Web Service Discovery) was designed and implemented.The point of this work is the algorithm designing and implement of ontology mapping processor in the framework.The similarity of entity in ontologies was computed by three different levels(linguistics,structure and instance).During the calculation of similarity based on linguistics,a method of construting a description of document was proposed and the method of edit-distance was adopted,besides that, WordNet dictionary was also introduced.During the calculation of similarity based on structure,an approach of adding structural characteristics information for each class node was proposed on the basis of the propagation theory of node similarity in SF algorithm.During the calculation of similarity based on instance,utilizing statistical methods to estimate the probability distribution of the concept instances of ontology, then calculate the similarity of entity in ontologies by Jaccard Similarity algorithm.In the process of similarity aggregation,a strategy of weight allocation based on coordination value was referenced,and reached the final similarity of ontology element pairs,then achieved the mapping between heterogeneous ontologies.Eventually,the validity and the accuracy of mapping algorithm was proved by OAEI2008 Data Sets, and the feasibility of the framework of Service Discovery was demonstrated by a service case.