Differentially Private Social Network Data Publishing

Abstract

Online social networks provide an unprecedented opportunity for researchers to analysis various social phenomena. Data collected by these networks are normally represented as graphs, such as connections among friends, which contain many sensitive individual information. Publishing these graph data without a proper privacy model may violate users’ privacy. In this chapter, we present two ways to achieve private social network data publishing using differential privacy: Node differential privacy ensures the privacy of a query over two neighbouring graphs where two neighbouring graphs can differ up to all edges connected to one node. And edge differential privacy means adding or deleting a single edge between two nodes in the graph makes negligible difference to the result of the query. However, existing works on differentially private graph data publishing only work properly when the number of queries is limited, as a large volume of noise will be introduced when the number of queries increases. A method called graph update method is then presented in this chapter to solve this serious problem. The key idea of the method is to transfer the query release problem into an iteration process, and update a synthetic graph until all queries have been answered. Compared with existing works, the graph update method enhances the accuracy of query results, and the extensive experiment proves that it outperforms two state-of-the-art methods, the Laplace method and the correlated method, in terms of Mean Absolute Value, showing that the graph update method can retain more utility of the queries while preserving the privacy.