Advanced Unsupervised Anonymization Technique in Social Networks for Privacy Preservation

We study the problem of anonymizing social network where the network data is split between several data holders or players. In such setting, each player controls some of the nodes and he knows only the edges that are adjacent to the nodes under his control. The goal is to provide anonymized view of entire network with respect to two scenarios. Scenario 1, where All players know the identities of all nodes but each player needs to protect the information from other players is the existence or non -existence of edges adjacent to his nodes. Scenario 2, each player needs to protect the identities of all nodes under his control along with the existence or non-existence of edges adjacent to his nodes. We start the study with sequential clustering algorithm applied to centralize and scenario 1 of distributed setting. Then we extend the algorithm to scenario 2 which is the most complicated part according to previous studies. Finally we conclude by outlining the future research proposals in that direction.