| In service-oriented applications, service integrators select the existing services and compose these services to provide the value-add service.Web service is a concrete realization of service-oriented architecture (SOA). The large numbers of Web services poses a challenge to service discovery. To discovery services through UDDI is based on keywords and simple classification.This mechanism can not meet the needs well,so it is one of the key issues to discover Web services quickly and accurately in order to meet user needs. For existing Web service discovery mechanism inefficiencies and the precision rate is not high, we proposed an ontology-based semantic Web services cluster and discovery framework.Firstly, we proposed the ontology-based Semantic Web service description model. By OWL-S description of semantic Web service, through Service Profile information of semantic Web services description, we get the Web services function information, such as service name, service description, input of services and output of services.Through the Service Model, Web service process description can be got.We use Petri nets to describe the execution order of Web service process .Secondly, we defined the similarity measure of Semantic Web service and analyzed the clustered Web service.Then we defined the concept semantic similarity algorithm based on domain ontology and the similarity of the process based on Petri net. It is the basis to cluster the Web services applications. Then we proposed the description of the gathering Web services.Finally, we proposed Semantic Web discovery method based on clustering Web services and done the experiment. In this method, Web service discovery process are divided into two stages,that are Web services categories matching and Web services matching. First filter Web services using the needed service description, then match the Web service based on semantic similarity. Experiments show that the proposed Web service discovery method can effectively improve the efficiency and accuracy of Web service discovery. . |
