Classification of Communal Blogs in Social Media Platform with Malicious user Content Detection

After crises/disaster events, millions of microblogs/social media websites include not only data regarding the current crisis, but also feelings and thoughts from the public. While most of the previous research has concentrated on collecting contextual data, it focuses on a specific non-situational category of tweets, i.e. social tweets, that attack other religious or racial groups in offensive messages. To create a classifier to differentiate between communal and non-communal tweets that performs much better than existing approaches. Ironically, a large number of group tweets are published by popular users, most of whom have to do with media and politics. In addition, users who post community tweets form strong, social network related communities. So there is a need to classify such posts from websites for security purposes. In this paper, a new method is proposed which is not event driven, to characterize the communal microblogs, this will be helpful in any disaster event. Also a procedure is built in which malicious URLs are also detected in such scenarios for security purposes.