Identification of markers and artificial intelligence-based classification of radical Twitter data

Purpose Social networks (SNs) have recently evolved from a means of connecting people to becoming tool for social engineering, radicalization, dissemination propaganda and recruitment terrorists. It is no secret that the majority Islamic State in Iraq Syria (ISIS) members are Arabic speakers, even non-Arabs adopt nicknames. However, literature researching subject deals with non-Arabic languages. Moreover, features involved identifying radical content shallow search or classification terms common daily chatter among region. The authors aim at distinguishing normal conversation, influenced by role religion plays life, terror-related content. Design/methodology/approach This article presents authors' experience results collecting, analyzing classifying Twitter data affiliated ISIS, as well sympathizers. used artificial intelligence (AI) machine learning algorithms categorize tweets, terror-related, generic religious, unrelated. Findings report accuracy K-nearest neighbor (KNN), Bernoulli Naive Bayes (BNN) support vector (SVM) [one-against-all (OAA) all-against-all (AAA)] algorithms. achieved high F1 score 83\%. work this paper will hopefully aid more accurate Originality/value In paper, collected analyzed thousands tweets advocating promoting ISIS. identified many markers keywords characteristic ISIS rhetoric. applied text processing AI techniques classify into one three categories: non-terror political news unrelated data-polluting tweets.