Research Of Privacy Protection Based On The Weighted Social Networks

With the rapid development of the Internet technology, the social networks are also accompanied by a rapid growth. Some well-known applications appeared, such as Facebook, WeChat, and so on. Large amounts of social network data are produced by using these social networking applications. And these data inevitably contain a large amount of personal privacy information. Therefore, in order to avoid disclosure of privacy, the data holders need adopt privacy protection polices before these data are released. The prior researchers mainly considered privacy protection of social networks which extend from the relationship data. However, there are still some problems in the research of privacy protection of social networks. On the one hand, there are not some suitable Chinese datasets for testing the validity of the privacy algorithm. On the other hand, most of the social network data are abstracted into simple graphs without weights, but the edge weights may be used as the attacker’s background knowledge, resulting in privacy leakage. Finally, most methods only consider the single sensitive property. In combination with the three aspects of content, this paper puts forward a algorithm to improve the strategy, the specific research results are as follows:(1) C-DBLP data recompiled. C-DBLP data is a collaboration network dataset which from WAMDM laboratory. In this paper, in order to compute edge weights for C-DBLP data, we obtain the published paper information by crawling from the original web. Then we compute the number of collaboration paper and the number of authors in each paper. Finally, we obtain the C-DBLP weighted datasets by integrating these data.(2) A generation method of weighted based on addition of edges. In real social networks, the weight can indicate tightness between two individuals of social relations. The weight may be as attackers’ background knowledge to re-identify the target individual and lead to loss of privacy. Therefore, in this paper, we consider protecting the weighted social networks from weight-based attacks and propose K-weighted generalization anonymity. This method combines K-anonymous with generalization method to ensure the security of the social network data when it is published. In order to ensure the higher validity of privacy protection, this paper introduces a concept of the weight difference to reduce the modification for weight graph.(3) A privacy protection method for multi-sensitive attributes based on weighted networks. We propose K-weighted L-multi-sensitive anonymity based on weighted social networks, and this method contains quasi-identifier and sensitive attributions at the same time. We make attribution generalization tree for each attribution according to the attributions of the real meaning. Our algorithm prevents targeted individuals from re-identifying based on quasi-identifier and realizes L-diversity for sensitive attributions.