Based On Clustering And Prediction Of The Semantic Web Service Discovery Research

In recent years, more and more service-oriented architecture SOA areapplicated by enterprise. Web services are to implement SOA the mostappropriate technology. But with the increase in Web services, service registry isan ever-growing, so how mass Web services efficiently and precisely to meetuser demand for services is becoming more and the greater the difficulty. Mostof the previous research Web service matching mechanism is based on thekeyword syntax match, at odds due to the lack of semantic information, makinghuman-computer inter-action, while in the service found that recall accuracy isnot very satisfactory. The purpose of the Semantic Web is to extend the currentWeb, so that all of the information has a wealth of semantic information in thenetwork, a computer can be enough to identify and deal with. Therefore, basedon semantic Web services is becoming a research hotspot.Web service discovery problem of inefficiency and accuracy is not high, wepropose a prediction based on fuzzy clustering and association rules semanticWeb service discovery method.First of all, the use of improved Fuzzy C-means clustering algorithm, fuzzyclustering preprocess the service in the service registry. With the domainontology modular thinking, there is a certain similarity of Web services inservice registry organization of fuzzy clustering. Service discovery, the servicerequest description and clustering similarity, filtering out irrelevant services,services to reduce the search time and improve the efficiency of servicediscovery; same time in order to improve the accuracy of service discovery,fuzzy clustering, all four functional parameters of services as a clusteringparameter clustering preprocess.Secondly, in the field of data mining association rules mining in SemanticWeb Services found that the potential relationship between the service requestfrom the user logging mining, the forecast will be requested service,recommended to the user, thusreduce service time and increase the efficiency of the service discovery. The same time, in order to improve the accuracy ofmining association rules, before digging the service log clustering preprocessbased on user context information, which can improve the accuracy of theService forecast found.Finally, on the basis of these studies, we designed and implemented asemantic Web service discovery model based on fuzzy clustering and predictionwere collected from the education, health care, food, travel, communication,economic and military seven field a thousandof the service as a test object, andverify the feasibility and effectiveness of the proposed method found in thismodel.