Privacy Protection on Social Network Data Using Anonymization Methodology

Privacy is amongst the major concerns when publishing or sharing social network data for social science research and also business analysis. Recently, researchers have developed privacy models just like k-anonymity in order to avoid node reidentification through structure information. However, even though these privacy models are enforced, an attacker can always have the capacity to infer one’s information that is personal if the list of nodes largely shares the identical sensitive labels (i.e., attributes). To put it differently, the label-node relationship just isn't thoroughly protected by pure structure anonymization methods. We present privacy protection algorithms which facilitate graph data to become published within a form in a way that an adversary who possesses information regarding a node's neighborhood cannot safely infer its identity as well as sensitive labels. To the present aim, the algorithms transform an original graph in to a graph during which nodes are sufficiently indistinguishable. The algorithms are created to accomplish that while losing only small amount information and even though preserving the maximum amount of Utility as possible.