Research On The Dynamic Friend Matching Based On Privacy Preserving In PMSN

With the wide application of smart phones and smart terminal, on line social networks also have rapidly developed, all kinds of social activities began to produce, people also faced some problems when they enjoy the funny applications which were brought by online social networks, for example, smart phones and smart terminal will have energy consumption and network restrictions problems, aiming at these shortcomings of online social networks, PMSN(Proximity-Based mobile social networks) can not only solve these problems, but also meet the user’s actual demand, it has become the hot research topic recently.Firstly, this paper introduced two kinds of typical methods of PMSN, privacy profile matching method based on trusted third party and distributed privacy profile matching, and describes the research status of two methods and existing problems.Secondly, according to the disadvantages of current privacy profile matching methods based on trusted third-party, this article proposed distributed calculation methods. It has high computational overhead and cannot meet the actual demand of users, aiming at two problems of existing distributed calculation methods, this article proposed ahead filtration protocol method. This method decreased the computing cost by reducing the number of users that need to match privacy profile. To better meet the user’s actual demand, this article used the correlation coefficient method to calculate the similarity.Then, in order to handle the computing cost problems of distributed computing method and the security problems of privacy profile matching method based on the trusted third-party, this paper proposed the double servers computing method based on the trusted third party. Double servers are anonymous server and computing server, the computing server has the role to compute similarity between the initiator user and candidate users, and the anonymous server is used to protect the users’ privacy of identities, it can not only decrease the computational cost but also ensure the security of users’ privacy by the method of double servers.Finally, experiments were carried out in the MATLAB environment for the real data sets. And compared with the typical methods, this text analyzed the performance of the two algorithms presented.