A Personalized a (d, k)-Anonymity for Social Network

Abstract

Privacy protection of Social Network becomes a more and more serious concern in applications. The common way to protect privacy is to use k-anonymity in data publishing of social network. In this paper, we study the theory of k-anonymity of social network, use weighted lower triangular matrix to represent the relationship among the nodes of social network, and propose a personalized a (d, k)-anonymity mode. The a (d, k)-anonymity algorithm experiments prove that it can make anonymous nodes of social network effectively resist d-neighborhood attack, and structure attack, at the same time, make the node information availability according to the personalized weight parameter a. The algorithm is usable, effective and efficient comparing with traditional (d, k)-anonymity algorithm.