The Research Of Trustworthy Service Selection Based On Community Technology

Service selection is a process that service users select required service from the servicecandidates which are function-equivalent with different non-functional properties. With rapiddevelopment of Information Technology, candidate services spring up overwhelmingly. How torealize effective and trustworthy service selection has become one of the hot topics in the researchfield of services computing.Traditional service selection methods need search services in large candidate services pool,which would affect the efficiency. Service selection methods based on collaborative filtering usesimilar neighbor’s service recommendation to avoid this deficiency, but these methods ignore theimpact of service popularity characteristics and time effect on the similarity calculation ofrecommenders; neither do they pay much attention on the trustworthiness of the recommendationusers, which badly restrict the accuracy of similar neighbors. Besides, these methods also ignore thetrustworthiness of QoS during the normalization process and not considering the specification ofusers’ requirements, only rank services based on QoS value simply, which badly restrict thetrustworthiness and accuracy of selection result either.To address these problems, this paper proposes a trustworthy service selection method basedon community technology. It improves traditional similarity calculation method and employs trustrelationship between recommenders. With the integration of similarity and trust degree ofrecommenders, trustworthy community is set up and this paper uses community rules to optimizecandidate services. By adding the process of trustworthy normalization of QoS and using prospecttheory to analyze the relationship between service quality and users’satisfaction, this method finallyprovides service requestor with accurate and trustworthy selection result. The simulation andapplication example demonstrate the feasibility and effectiveness of the proposed method.