Semantic Web Service Discovery And Composition Based On Social Network

Semantic web service is based on service oriented architecture (SOA). Semantic web service has a wide range of applications in e-commerce and integration of enterprise application with its advantages of self-contained, self-describing and cross-platform etc through service reuse. And it has become the focus in the areas of industry and academic. Service discovery mainly refers to precision rate and recall rate, while service composition focuses on automatic composition of the services.As the social network can make full use of the context of the relationship between network entities to find the network entity, the researchers combine semantic web service with the social network, and put forward the social network of semantic web service to improve the efficiency of semantic web service discovery. However, the combination of semantic web service and social network also brings new problems:how to build the social network of semantic web services; how to find a semantic web service in the formed service network; how to optimize the formed network of semantic web services; how to combine services in the formed network, etc. This paper focuses on these issues to research the application of the social network in semantic web service. The innovations are as follows.(1) The directed-weighted social network of semantic web service is proposed. The existing social networks of semantic web service are mostly undirected and unweighted so that this method can only research the social networks of semantic web service from a qualitative point of view. To solve this problem, the semantic web services and their relationships are formalized as the directed-weighted service network. This method, in which the social network of semantic web service is established, is mainly reflected in that the set of semantic web services is regarded as the node, the connection between service sets is regarded as the directed edge, and the strength of the connection relationship is regarded as the weight. Meanwhile the interaction mechanism of domain ontology library, web service requestor, web service provider and web service registry is proposed to update the service network. The result of the experiment in the goods service composition process shows that the established service network through this method can better describe the characteristics of semantic web service.(2) The hierarchical clustering algorithm based on the density is proposed to build the social network of semantic web service. It is difficult to determine the number of clusters in the clustering results with the existing clustering method for semantic web services, so that more parameters are needed to be set manually during the clustering which results in that these parameters are applicable to a particular data set. To solve this problem, the hierarchical clustering method based on density and interconnection is presented. The similarity of semantic Web services is used to cluster, so that the social network nodes of the semantic web services are formed. It is supposed to refine the classification of semantic web services by controlling the density and interconnection of the nodes to adjust the number of the semantic web services in the nodes. Meanwhile the two-stage inverted index method is adopted to improve the query efficiency of semantic web service in the service network. The two-stage inverted index method with ontology-service and service-node are employed to discover the semantic web service. Using OWL-TC4data sets, the experimental results show that the improved hierarchical clustering method increases the accuracy of clustering semantic web services, and the two-stage inverted index method reduces the number of querying semantic web services, so that the query efficiency is improved.(3) The method based on the cost of compensation is proposed to optimize the service network structure dynamically. Most of the existing research methods only consider the service network which is composed of semantic web services in a moment without considering the structure changes of the service network at different times, so these methods are not adaptive to the dynamic nature of semantic web service. To solve this problem, the dynamic optimization method of the network nodes based on the time series is proposed. The method is supposed to describe temporal dynamic and static moment of the service network with time series, which means that the service network is static at every moment of time series and changes dynamically with time series. The compensation cost of semantic web service refers to the price to pay for realizing the required functions in the case of update or failure of semantic Web service. By comparing the compensation costs before and after updating the network nodes of semantic web service, it is supposed to determine whether the network nodes need to be updated, so that the nodes which semantic web services belong to are determined dynamically. Taking QOS data sets in different periods for experiment, the results show that the method can effectively reduce the times of changes in the network structure, obtain the stable service nodes, and improve the query efficiency of the web service.(4) The method based on the service network structure is proposed to verify the composition process of semantic web service. Most of the existing composition processes of semantic web service employ formal method for verification, and the efficiency of verification for service composition get lower along with the increasing number of semantic web services. To solve this problem, the verification method of the web service composition based on the reachable probability of service nodes is proposed. The nodes which semantic web services belong to in the service composition process of service network are taken as the research object. The composition process of semantic web service is turned to the reachable path of service nodes. By calculating the attainable probability of two service nodes based on the weight of the directed edge and combining with the network centrality and the critical path, it is supposed to determine whether the service composition needs to be verified. Taking simulation data for experiment, the results show that the method can effectively reduce the times of verification and improve the verification efficiency for service composition.In summary, the social network is used to study the discovery and composition of semantic web service in this paper. As a result, the method has obtained certain research achievements, and has important theoretical significance and application value to promote the research of semantic web service using the social network.