Prevention of information leakage by modulating the trust uncertainty in Ego-Network

Leakage of personal information is a problem with free usage of social networks. In the face of enhanced connectivity reaching out to friends of friends, it is becoming increasingly difficult to prevent unwanted users from seeing personal posts. Various automated means of abatement measures are being taken in this regard and the present work is an attempt in that direction. Conventionally, the trust value between two mutual friends is computed based on attribute matching among them. In this work, the trust is computed by blending with the degree of the target nodes to arrive at a modified trust metric based on the viewer-ship decision which is taken by user. It is explored that as the emphasis of the blending factor shifts from the conventional trust towards the degree factor, the uncertainty in trust values increases while the viewer count of a post keeps reducing simultaneously. It is proposed that the blending factor may be tuned at the crossover point in order to suit the needs of the user in these two respects. Detailed testing and statistical analysis of proposed scheme has been conducted on representative data of Facebook available in public domain.