Understanding Network Characteristics of Spam Users in Social Media

Timely identification of spam users, usually regarded as users creating and spreading false information in social media, has experienced a resurgence of interest. Besides the linguistic and behavioral cues, the network among users has become an informative yet robust data source towards spam users detection. However, the characteristics of spam users’ behavior on networks have not been closely examined. Along this line, our study selects two datasets from e-commerce review system and social networking platform for systematically exploring network characteristics of spam users. We perform the network analysis from three aspects: homophily of edges, centrality of nodes and cohesive groups. Our analytical results yield several conclusions: (i) the homophily property is always satisfied on the e-commerce co-reviewing network, but it does not hold on multiple relations of social networking platform; (ii) spam users are likely to be on the periphery of the network on e-commerce platform, but on the innermost network core on social networking platform; (iii) spam users on e-commerce platform tend to form cohesive groups that are quite scarce on the social networking platform.