Links reconstruction attack: Using link prediction algorithms to compromise social networks privacy

The explosion in the use of social networks has also created new kinds of security and privacy threats. Many users are unaware of the risks involved with exposing their personal information, which makes social networks a “bonanza” for identity thieves. In addition, it has already been proven that even concealing all personal data might not be sufficient for providing protection, as personal information can be inferred by analyzing a person's connections to other users. In attempts to cope with these risks, some users hide parts of their social connections to other users. In this paper we present “link reconstruction attack”, a method that can infer a user's connections to others with high accuracy. This attack can be used to detect connections that a user wanted to hide in order to preserve his privacy. We show that concealing one's links is ineffective if not done by others in the network. We also provide an analysis of the performances of various machine learning algorithms for link prediction inside small communities.