Privacy preserving graph publication in a distributed environment

Abstract

Recently, many works studied how to publish privacy preserving social networks for "safely" data mining or analysis. These works all assume that there exists a single publisher who holds the complete graph. While, in real life, people join different social networks for different purposes. As a result, there are a group of publishers and each of them holds only a subgraph. Since no one has the complete graph, it is a challenging problem to generate the published graph in a distributed environment without releasing any publisher's local content. In this paper, we propose a SMC (Secure Multi-Party Computation) based protocol to publish a privacy preserving graph in a distributed environment. Our scheme can publish a privacy preserving graph without leaking the local content information and meanwhile achieve the maximum graph utility. We show the effectiveness of the protocol on a real social network under different distributed storage cases. © 2013 Springer-Verlag.