Identifying terrorist index (T+) for ranking homogeneous twitter users and groups by employing citation parameters and vulnerability lexicon

Abstract

Twitter, one of the important social media also acts as substrate to mix enzymes of actions and reactions of public sentiments including terrorist activities, either implicitly or explicitly. The identification of terrorist or crime related texts from Twitter is difficult because of the presence of such implicit and implied code-mixed hints that are solely shared among different users within a few groups. In the present task, we have developed a framework that deals with Twitter data and provides varieties of dataset useful for tracking various terrorist activities. The system also identifies the terrorist indices of different users as well as their groups. Compared to the available citation indices (e.g., h-index or i10-index), this newly proposed index takes into account the normalized parameters of citation and seed as well as code words related to crime from vulnerability lexicon. Finally, this index is used for ranking the Twitter users and groups and it is observed that our proposed ranking algorithms have performed reasonably better when the algorithms encapsulate the roles of citation and vulnerability parameters together.