ABSTRACT
Online radicalization is hazardous to society. Social networking has been utilized as a significant hotspot by the extremists for directing radical activities. Twitter has gained much attention among the radical groups due to its popularity among users, ease of communication, and having short text messages containing noise thereby making them difficult to detect. Most of the radical Twitter accounts have been suspended by the government law enforcement agencies but still, some radical activities are going on through unrevealed accounts that are to be detected yet. A comparative analysis of the techniques used by different authors and also the challenges in detecting online radicalization are discussed. Further, an effort is also made to detect radicalization over Twitter using a set of semi supervised machine learning techniques.
