![]() | Only 14 pages are availabe for public view |
Abstract Social media platforms are becoming more widespread and popular than ever as people use them to communicate with others, express what they feel, their daily routines and announce their opinions on world events. Radical groups may exploit its features of self-promotion, relationship building, news and information posting to disseminate their ideologies, fake news, terroristic propaganda and spreading hate speech In particular. As some violent radical groups such as ISIS and Al-Qaeda have developed well designed propaganda strategies that enables them to recruit more members and supporters all over the world using social media facilities. So it is crucial to find an efficient way to detect the violent-radical accounts in social media networks. Gaining insight of these Radical groups methods and techniques to spread their ideologies and firmly analyzing their strategies in recruiting new members reveals the suitable countermeasures to their destructive influence. In this thesis, a new-labeled Arabic ISIS related tweets and writings dataset is introduced. The method of gathering, annotating, properties, and statistics of the dataset will be discussed. A set of benchmark experiments using different types of machine learning algorithms, features and preprocessing to reach the best detection accuracy will be presented. Finally an intelligent system that autonomously detects ISIS online community in Twitter social media platform is proposed. |