Research On Web Services Selection And Recommendation Method Based On QoS Metrics

Web service has become technical supporting power for various fields such as military,medical care,education,and business,and it has injected invaluable development forces into various industries with various service combinations of different functions creating increasing new value-added services.However,with the continuous development of the network technology and constant changes in the scenario of application,Web service,based on ubiquitous use,keeps adapting to the new network environment and coping with all kinds of new application challenges.The demand for Web service is also shifting from function realization to QoS(Quality of Service)improvement.Although users have always paid high attention to the QoS's description of the Web service property,their perspective of the concern has changed from the past.The QoS property study has turned from the simplex and details to the multi-angles and integrality;and it strives to make improvement targeting the application traits of Web service selection and combination in different network forms.This paper addresses the poor credibility of QoS data,low selection efficiency of Web service and inaccurate service recommendation in the prevailing mobile network with expanding application of Web service,and it has achieved the following results:1.Web service reputation measurement method based on user feedback evaluation.Credibility,as an attribute of service quality that can reflect the overall quality of Web services,has become an important basis for people to choose services.It simplifies people's evaluation of Web service quality to some extent.Good credibility is a prerequisite for people to choose a Web service.However,based on people's concern about the credibility,a service provider may take advantage of unfair competition means to improve the probability of their published services;they may personally give themselves high grade and opponents malicious low grade or hire others to do so.These untrustworthy evaluations will cause great deviations from the actual QoS value,and then affect the accuracy of Web service selection.In view of this,this paper proposes a method for measuring the credibility of Web services based on user feedback,mainly from two perspectives--timeliness and comprehensiveness.By comprehensive evaluation of direct and indirect reputations and the consideration of time weight,the method weakens the impact of malicious feedback on credibility.The experiment verifies the advantages of this method in the credibility measurement accuracy.This method provides reliable guarantee for the authenticity of the data in subsequent selection and recommendation of Web service.2.Research on Web service selection method based on QoS uncertainty.The QoS value of a Web service is not always stable within a certain range,showing highs and lows even largely deviating from the normal range.The reasons behind it vary.In addition to trustworthiness,good stability is also a prerequisite for guaranteeing the high performance of a service portfolio in the process of service composition and selection.Therefore,a highly accurate and reliable Web service QoS measurement method will significantly improve the quality and performance of the entire composite service.To solve this problem,this paper comes up with a Web service selection for QoS uncertainty,which calculates the QoS uncertainty by using the coefficient of variation to filter out poorly stable services and optimizes the service selection with a mixed integer programming algorithm.The experiment verifies that the method can optimize the Web service selection in terms of running time and reliability.3.Research on Web service recommendation method based on users mobility.In the environment featuring ubiquitous mobile Internet,user context often changes when people use Web services.This ever-changing user context can cause users to generate different QoS values when using the same service.In this case,there is a significant difference in the experience of the same service,but even if the network environment changes,people also want to get the same kind of Web services that the high quality of service is in line with their own demand.In order to solve the above problems,this paper proposes a method of Web service recommendation based on QoS prediction.In the process of QoS prediction,the changes of user context are fully considered,and the similarity comparison is used to supplement the QoS in the absence of QoS,and to recommend services that meet the users' requirements.Experiments show that the QoS predicted by this method has higher accuracy and relatively less time consumption.