Research On Method Of Personalized Privacy-Preserving Data Publishing In Social Network

With the development of network and information technology, the internet has been into every aspect of our lives. Especially, with the development of the Web2.0technology, more and more social network products have played an important role in our daily life. Social networks, such as Qzone, Renren, Pengyou, Weibo, Facebook and Twitter, provide us convenient service of chatting with friends and sharing information, thus make lots of people take part in and have been one of the most rapid developments of network application.A sharp increase of social network data is becoming one of the most valuable resources of the internet. It is used for a variety of social science research and business analysis, such as the community behavior analysis, the social structure analysis and the public opinion monitoring, etc. The social network data-centered analysis has become an important aspect of the academics and business, and the publishing of the social network data has become a significant application demand. But publishing the social network data may generate individual privacy disclosure risks, since the social network data contains a large amount of personal information. How to publish the social network data while preserving the individual’s privacy has become one of the hot spots of current research.In recent research about social network data publishing, social network data can be represented as a graph, in which nodes and edges correspond to social entities and social links between them, respectively. The social network data and the traditional table data have difference, since the individuals are independent of each other in the existing privacy preserving research of table data, but the individuals have connection between each other in the social network. Therefore, the methods for the table data publishing cannot be used for the social network directly. At present, researchers have developed lots of privacy models and anonymous techniques to prevent re-identifying of relevant information of nodes through structure information of social networks, but the anonymous technology needs to change the structure of the graph, especially for the social network that has a small number of nodes with high degree distribution (such as power law distribution), the existing anonymous technology needs to add lots of edges and leads to serious data distortion. One of the most important concerns in publishing social network data for social science research and business analysis is to balance between the individual’s privacy protection and data utility.Through the investigation and analysis of the social network, in fact in most of the social network, the users’ privacy requirement and sensitivity are differentiated, that is, the nodes’ privacy preserving requirement in social network are diversified and personalized. There is no doubt that exerting a universal privacy preserving strategy for all nodes will cause some extent of "excessive protection", and then bring unnecessary data distortion. Motivated by this, the research direction of this paper is centered on the model and method of social network data publishing with personalized privacy feature. The main work in this paper is as follows:First, we research the current situation of privacy-preserving technology in social network data publishing. We discuss the personalized problem in social network data publishing. We describe the privacy model for personalized privacy preserving in social network data publishing, which contains the attackers’ background knowledge for privacy attack and the evaluation standard of privacy preservation.Second, we analyze two anonymity methods deeply:k-degree anonymity, k-degree-1-diversity anonymity. Based on the actual privacy protection demand of edges’ sensitivity, we propose (k,1,p)-anonymity model. And according to this, we combine the individuals’ personalized privacy preserving requirement and put forward a new personalized (k,l,p)-anonymity model and algorithm to reduce the distortion extent of the data in the privacy processing of data.Last, based on the proposed algorithm and model, we implement a prototype system which is about personalized characteristics of social network data privacy protection. We also conduct experiments on some real-world datasets to evaluate the practical efficiency of the privacy preserving methods, which not only guarantee the extent of data privacy protection, but also improve the utility of the data.