Research On Web Services Semantic Relationships Mining And Service Composition

With the advent of the era of big data, web services technology has been furtherdeveloped. There has been a large number of stable and easy to use Web Service, how willsingle, limited functionality of web services combine to provide users with comprehensive,meet the personalized needs of composite service has become a current research focus. Inthis paper, the service class as a research object, the service relationship as a support, byresearching the service semantic relationships between service classes to achieve theautomatic service composition of web service.(1) Study the semantic relationship between web services. Defines the semanticrelationship between web services, and finished the service library division through serviceclustering. Take the QoS attributes as filter criteria, through screening atomic service, andestablishing the index of service QoS dimension preference to complete service classoptimization. This paper is based on the service class, improves the granularity of servicerelationship mining, and proposes web services’ semantic relationship mining algorithm, toprovide semantic relationship support for the automatic service composition.(2) Study the automatic service composition method of web services. The servicecomposition process is divided into two phases: the service classes’ composition andservices’ binding. Service class composite phase used planning algorithm to complete theconstruction of the semantic relationship graph, and then the service composite mapped asthe shortest path problem to solve service class composite scheme. This paper also proposesa personalized service selection method to complete the specific service binding.(3) Study the service selection method of web services. To meet the personalized needsof users, this paper proposes a service selection algorithm based preferencerecommendation and QoS. By calculating the similarity of evaluation preferences betweenservice requester and historical evaluation users to get the service requester’s suggestedusers. Then calculate suggested user the corresponding candidate services’ recommendeddegree, and finally combine service requester’s QoS preferences to complete thepersonalized web service selection.Through a series of experiment to verify the service class semantic relationship miningalgorithm, service composition algorithm and service selection algorithm. The results of theanalysis show that the proposed algorithm is feasible and effective. Research results caneffectively solve the atomic service and its semantic relationships massive growth problems,and provide personalized and high-quality composition services to further improve the usersatisfaction of service request.