Research On Semantic Similarity Measuring Method For Web Services

Service oriented computing has been a new computing paradigm, service composition becomes an important software development method. Semantic Web is one of the core technologies in future networks ,it presents more complex concepts and relations . Web has been translated from a platform of information sharing to service integration platform. Now, Web services are archiving dispersedly around the internet.The various complex service requisite can be met by service composition, that is to say, by combining all kinds of simple web services from web. Service discovery is to find a usable requisite Web service from web for Service compositions . One important and critical technology in service discovery is the similarity measurement method for web service .This thesis analyzes the different computational methods of existing Web service semantic similarity and the advantages and disadvantages of them. Based on the distance model and information content model, this thesis puts forward a semantic category description model for Web service (SCDM4WS) and extends the semantic information in the Web services description,which build semantic foundation for similarity measurement of Web service .Furthermore, it gives a definition of the concept of the semantic distinguish accuracy.Secondly, based on SCDM4WS model, this thesis brings forward a measurement method SSMMWS of Web semantics similarity which can argument the semantic distinguish accuracy. This method considers synthetically various factors including distance, depth and density which affect semantic similarity. This method can distinguish different density of brother nodes well in the distance model;It can distinguish different depth of nodes well in the information model;and it avoids the time-consuming and space-consuming problems resulting from the text data set statistic processes.Thirdly, this thesis adopts the Tag technology in service semantic similarity calculation, and puts forward a semantic similarity measuring method SSMMWST based on Tags for Web services . The method uses the Tag semantic information to compute semantic similarity of Web services.Thus this further improves the calculation accuracy of Web service semantic similarity.Finally, This thesis designs and achieves the experiments, and verifies the feasibility and accuracy of the above method. We adopted the data sets which is provided by German Research Center for Artificial Intelligence to experiment. we compare the Roy method, Resnik method,Cosine (Cos)method,Extended Jaccard (EJ)method with the method presented in this thesis and analyze and discuss the experimental results. Experimental results show that the method of the thesis is feasible and efficient.