Identification of Sybil attacks on social networks using a framework based on user interactions

Abstract

The popularity of modern social networks has rendered social media platforms vulnerable to malicious activities. One such activity is a Sybil attack, in which a single entity emulates the behaviors of multiple users and attempts to create problems for other users and a network itself. This problem has prompted researchers to develop several techniques for preventing Sybil attacks, but in most cases, the efficiency assumptions that underlie proposed methods are not oriented toward reality. The current study puts forward an efficient framework for identifying Sybil attacks. The highly precise framework is underlain by rational assumptions and detects attacks on the basis of the structural characteristics of social networks and the social interactions among users. We evaluate our proposed framework using both synthetic and real world social network topologies. We show that SybilUncover is able to accurately identify high precision rate. Moreover, SybilUncover performs orders of magnitudes better than existing Sybil detection mechanisms.