A Novel Technique to Characterize Social Network Users
Monika Singh, Divya Bansal, Sanjeev Sofat · 2016
In the present digital network, where people share critical information through online social networks on a daily basis, it becomes essential to identify security loopholes and vulnerabilities in social networks. Social spam has become inevitable with increasing number of users exploiting it for communication. Spam profiles have become dodgy for the social platform and beyond, since they pollute the network with insignificant and malicious information. This put an adverse impact on economy, politics, and society. In this paper, we contribute along various dimensions. First, we created a large dataset of genuine and spam profiles and exploited it for validation purpose. Second, we exploited trust based feature for anomalous accounts detection. Third, we proposed a unified framework to classify various types of Twitter users. Fourth, results of the proposed work have been compared with three existing state-of-the-art techniques showing the effectiveness of the proposed technique based upon promising feature selection.