Study On Greedy Algorithm Based Privacy-Preserving Approach In Social Networks

In recent years,with the popularity of the network,Internet has access to all aspects of human life. Especially with the rise of Web 2.0,various social networking sites have sprung up.Due to the huge number of users,people can easily get a lot of personal information from the network.If do not use the information correctly may pose a serious threat to personal privacy and information security.Therefore, the privacy security of network users is crucial.In this paper,based on in-depth study of existing research results,we proposed a Greedy algorithm based privacy-preserving approach in Social Networks(GSNPP) and explored from the following aspects:1 Improving the existing social network privacy protection methods,establish Social Networks privacy-preserving model,using a combination of individual identity attribute and associated structures attribute to anonymization Social Networks.2 Using GSNPP algorithm to cluster the nodes in the social networks. The algorithm solved the problem of nodes selected in the anonymous process of the social networking and quantified the different types of information loss brought about by the anonymous process.The principles of clustering is:(1) Meet the dissimilarity between selected nodes and the nodes in the cluster is relative minimum.(2) Meet the loss of information of selected nodes is relative at least in the generalization process.(3) Each cluster should satisfy k-anonymity principle,that should contain at least k nodes.3 Propose a A personalized privacy-Preserving Approach,which introduce an importance weight parameters between data attribute information and structure attribute information,so that according to the the importance of different demand on different attributes of users,we can adjust importance weight parameters flexibly to balance the anonymous degree of different attribute information,and achieve personalized privacy protection.5 Simulation results demonstrate that the proposed GSNPP Algorithm based Privacy-Preserving Approach to test the feasibility and effectiveness.