Detection of Account Cloning in Online Social Networks

Abstract

© 2020 IEEE. This research proposes a framework to detect account cloning in online social networks. The main concept is to analyze user profiles, friend and follower networks, and user posting behaviors. The framework consists of 3 parts: Twitter Crawler, Attribute Extractor, and Cloning Detector. Twitter was used as a case study. Experimental results suggested that it was not easy to completely clone profiles and friend/follower networks of the victims. And even the attackers managed to do so or somehow fool other users, posting behaviors and writing styles could help distinguish between fake and authentic (i.e. done by the victims) posts. The average accuracy of classifying whether the posts were fake or authentic was 80%. Decision tree was found to yield the best classification performance.