Online Extremism Discovering through Social Network Structure Analysis

The activity of extremist organizations on the Internet is continuously growing with the increase of Web's usage for means of communication. Therefore analysis of radical members in social networks provides important information on how to prevent them propagate ideology and recruiting new members in the future. But nowadays terrorists often use confidential chats and private threads to communicate, thus it's quite hard to detect them using only the public messages they generate. In fact, it is usually known that some users of social networks are dangerous, another are innocent, and no information is available about the remaining users. In this paper, we propose an approach for detecting radical users of social network among unknown ones by analyzing their relationships and features as of vertices of social graph without usage of any information about text content they generate. We find that the proposed method is very promising and may be efficiently used for real-time monitoring systems and future terrorism and extremism research.