Research And Implementation Of WEB Service Recommendation System Based On Important Users

With the development of Internet technology,more and more Internet companies have become service providers.A mount of Web services which have similarity function and description are provided by them on various Web service platforms,which makes it difficult for customers to find their favorite services quickly and accurately.To overcome this problem,the Web service recommendation technique has been widely employed for personalized service recommendation recently.Although the service recommendation systems have shown their potential to solve the information overload,they are still facing many challenges,for example the problem of shilling attack.It is known that more data exists in recommendation system,the recommendation results are more accurate.So most of the systems are open to all of its users.This feature of openness makes the malicious users have the opportunity to inject fake information to upset the recommendation systems.And some users input data optional to effect the stability and precision of systems.So the recommendation systems appear to be vulnerable to shilling attack problems and useless data.To solve this problem,this paper,therefore,provides a Web service recommendation system based on important user group.The main work includes:(1)In this thesis,the background,research progress status and the research significance of the Web service recommendation systems are introduced.At the same time,we put forward the main research content of this topic.And the relevant technology of Web service and collaborative filtering and shilling attacks are discussed.(2)This paper analyzes the robustness of Web service recommendation systems in detail and studies the possible attack models in Web service recommendation systems deeply.And we also presents Web service attack models,like normal attack models and the Pareto Web service attack models and Web service attack models based on the geographic location.Furthermore,this paper accomplishes five experiments in the real dataset WS-DREAM.And the results show the effectiveness and rationality of these attack models.The Web service recommendation systems appear to be vulnerable to shilling attack problems.(3)This paper composes the recommendation algorithm based on important user groups(IMURS).It computes user importance by the quality of Qo S(Quality ofService),user action and user social relationship.The experiments on the datasets of WS-DREAM,Epinions and Mei Zu APP store show that the IMURS has higher accuracy and better robustness facing the shilling attacks.(4)This paper designs a simulation IMURS model based on the theory of software engineering and demonstrates the main functions of the system.