Participatory detection of identity theft on mobile social platforms

Abstract

Popularity of smart devices has led to increasing use of social networking services for various purposes. Despite its benefits, the ubiquity of social network services introduces vulnerabilities to malicious behavior such as Sybil attacks or identity theft. As the 5G Era leads to the convergence of social, wireless, and mobile networks by enabling synergistic interplay between these networks, it is possible to take advantage of mobile edge computing in the detection of compromised social profiles in mobile and online social network platforms. In this paper, we propose a framework for participatory detection of identity theft on social networking platforms which would exploit the computing power of the user equipment. The proposed framework empowers the connections of a user in a social platform to cooperate on the verification of a social profile. Through a proof-of-concept study, we show that the proposed framework can detect anomalous behavior in the social profile by having each connection work on a different feature subset without semantic analysis. Our numerical results show that if the initial matching threshold in a decision tree is set properly, compromised accounts can be identified by the mobile platforms of the connections without undergoing heavy central processing.