Trusted QoS Service Selection Method Based On Evaluation Classification

As so many services are involved in service computing,service selection is becoming more and more important.The feedback evaluation information given by service users is directly use into service selection indiscriminately,which may lead to a big deviation between the evaluation result and the actual situation.The existing method of service selection based on trusted QoS only carries out the screening of trusted users,and then uses all the evaluation of trusted users directly for service selection.In fact,some individual subjective evaluation habits of trusted users may also result in some deviation in service evaluation.The indiscriminately direct use of the feedback evaluation which exist deviations will have a greater impact on the accuracy of the service selection results.In order to solve all the above problems,a trusted QoS service selection method based on user evaluation classification is proposed in this paper.The method not only considers the credibility of the user feedback evaluation,but also takes into account the influence of the subjective evaluation habit of the trusted users on the evaluation of service feedback.First of all,the user evaluation is classified in detail according to the service attributes on the basis of the analysis of user evaluation habits.Discriminate and reject malicious users based on classification evaluation.In the meantime,the deviation evaluation of trusted users is marked,and the deviation degree from the evaluation data is further calculated to correct the deviating evaluation data of the trusted users,so that the corrected deviation evaluation tends to be reasonable and the reference value of deviation from the evaluation data will be higher.In addition,in order to improve the accuracy of service results,this article is based on the revised user evalution of similarity calculation for candidate services overall evaluation vector: obtain the mean vector each candidate services based on the revised user evaluation,establish the trusted user feedback evaluation weight of reference based on the user evaluation vector and the mean vector similarity,and then calculate the QoS comprehensive score of each candidate service and complete the service selection according to the personalized preferences of the target users.The effectiveness of this method has been verified by a large number of simulation experiments.