Research Of Privacy Protection For Personal Information And Relationship Of The Social Networ

As the internet emerges in the social life, Social Network Service (referred as SNS) as one of the products of the Web 2.0 technologys has become the hot research topic. The myth of the great increase of users and wealth are created by Facebook and other SNS applications again and again. People involved in the social networks share and publish more and more information. The exposure of sensitive information becomes a serious problem. This paper focuses on two aspects of privacy protection in social network: privacy protection of the personal information and the privacy protection of the relationships.The first research of this paper is about the privacy protection of the personal information. The models of the traditional tabular data for privacy protection like the K-Anonymity model, L-Diversity model and the personalized model are studied at first. Considering the individual needs of the privacy protection in the social network and the features in the release of personal information, the personalized (P,α, K) anonymity model is proposed in this paper. The hidden field P, hidden levelsαand the same number of identifiers K and other factors are considered in this model. It reduces the running time. The procedure of the anonyimization and the generalization is showed. How to calculate the loss and the disclosure of privacy information is discussed then.The second research of this paper is about the privacy protection of the relationships. The algorithms of the privacy protection in the release of the graph information are studied at first. The features of the social networks are similar to the graph information. Considering the major individual needs of many users in the social networks,based the three levels settings of the privacy protections, thefour levels model is proposed. This model allows the user to set different levels of privacy protection, and combines the privacy protection of the edge information with the privacy protection of the micro-data. There are four different levels, from low to high, the protection of the label information, the protection of the node degree, the integrated protection of the label information and degree, the protection of connected nodes. The main purpose of this approach is to provide different users with different services.Finally, a simulation platform is built in Delphi, and MySQL. This platform is built to achieve the personalized (P,α, K) Anonymity model. The running time and the information loss are analyzed over the experimental data. Then the model is proved properly for the privacy protections of the personal information in the social networks. By modifying the open source visualization of social networks called Vizster, the model is able to increase the function of adding nodes and edges, allowing the user to set four levels of privacy protections. Then it is proved properly for the privacy protections of the relationships in the social networks.