Research On Technology About Privacy-preserving For Cross System Personalization

With the gradual increase of the information resources on the internet, users'demands for personalized services are increasingly heightened. Cross system personalization by sharing the personalization information between different systems in a user-centric way, to obtain maximum leverage and reuse the information about users, thus enhancing the quality of personalization. However, due to the strange of the systems and the sensitive of the information shared, many users will not like providing the information about themselves. Therefore, how to make the information shared effective without revealing the users'privacy becomes the main issue in the development of cross system personalization.According to the contents of the information shared, the contribution of this paper contains three points as follows.Firstly, aimed at the users'profile shared across systems, considering the information may be eavesdropped or altered by the unauthorized third party during the process of the transmission in the network. In the framework of cross system personalization based on register center, adding the digital signature, identity authentication and multiple encryption technology, provide a hybrid encryption system to implement the data exchange across the systems. It can ensure that the secrecy of the data, the integrity of the data and the denying of the participant in the process of the transmission.Secondly, for the commercial competition or any other reasons, users are not wanted to provide individualized information. Based of the theory of secure multi-party computation and combine the scalar product protocol, an effective cross-system privacy preserving collaborative filtering recommendation algorithm is presented. The algorithm uses oblivious transfer protocol to prevent that the un-trusted third party and systems from colluding. It also can improve the accuracy of the collaborative filtering algorithm.Finally, realized the critical technology of the data exchange across the systems which framework is based on the register center. We also have given the experimental schemes for cross-system privacy preserving collaborative filtering recommendation algorithm. The algorithm has been proved to be feasibly by experiments in the aspects of time performance and accuracy.