Research On Privacy-Aware Recommendation Technology For Mobile Commerce

With the development of information and communication technologys, mobile commerce has entered the age of rapid development and the location-based service(LBS) has been used widely. However, “information overload” problem increasingly prominent; at the same time, in mobile application, users’ privacy protection in LBS has become a hot issue. Therefore, providing personalized recommendation service based on the protection of users’ privacy become a new challenge to solve to further development of mobile commerce.This paper mainly focuses on privacy problem and “information overload” problem in mobile commerce, studies the new model of mobile commerce applications sense of prvacy, privacy protection technology in mobile commerce and recommendation technology based on privacy protection. The major works of this paper include:(1)A privacy preserving mobile service recommendation framework based on cloud(CMR) is established in the research of privacy concerns in mobile commerce, privacy preserving technologies in the mobile environment and personalized recommendation algorithms. Then, the function of each module of the framework are analysed and the recommendation service process is described in detail.(2)According to the framework of CMR, a new privacy preserving algorithm named as EMAGAS is proposed, which features the construction of minimum initial K-anonymity sets, exchanging process and merging process. To show the availability of EMAGAS, operation method and the results of each major step through an example application are illustrated. Finally, based on a real road network and generated privacy profiles of the mobile users, the feasibility of EMAGAS algorithm are validated by experimentlly analyzing and comparing the performance of EMAGAS and P3 RN using the metrics including information entropy, dummy ratio, query cost, anonymization time. The experimental results demonstrate that EMAGAS has advantages over P3 RN.(3)Based on the framework of CMR, combing with bipartite graph, TOPSIS-attribute decision method and K-NN query technology, the personalized recommendation technology sense of privacy based on energy diffusion theory and TOPSIS is proposed. The feasibility of the algorithm is validated by example analysis.The main conclusions are drawn and the interesting points for future work are provided in the end of dissertation.