Identifying Sources of Misinformation in Online Social Networks

The importance of online social networks as a media for dissemination of news has increased in the last decade. The real time nature of the contents and the speed and volume of propagation have posed great challenges to assess the quality of information in an acceptable time frame. Collusion of users to spread false information and simultaneous spread of multiple false messages have made their detection a challenging task. In this paper we propose a methodology based on principles of cognitive psychology for detecting and monitoring sources who collude with each other to spread misinformation. We use social network as a social computing platform to classify sources as credible or non-credible based on the level of acceptance of their messages by other users and patterns of propagation. The proposed methodology could form a framework for an effective social media monitoring system. We have implemented our algorithm in the online social network 'Twitter'.