Analyzing Behavior of ISIS and Al-Qaeda using Association Rule Mining

Abstract

Social media have been exploited by terrorist groups to share their massacres plans, recruitment, buying weapons, or propagating their violent plans. Terrorist groups named ISIS and Al-Qaeda are the most active and well known for using social media to propagate their violent intents over online discussion forums. It becomes necessary to study the behavior of these terrorist groups over online social media. In this paper, we present association rule mining based approach to extract a feature set for terroristic groups named ISIS and Al-Qaeda. We used the Global Terrorism Dataset which contains systematic information on terrorist attacks worldwide since 1970. Entropy-based feature extraction technique is used to extract top features which are then further used to find association rules. Eclat (Equivalence Class Transformation) and Apriori algorithms are used to mine association rules from prepared data. Rules for ISIS and Al-Qaeda are computed separately, and are then further classified using machine learning classification algorithms. Our research contributes to the smart and novel application of data mining algorithms and computational intelligence to study the behavior of the most popular and active terrorist groups over social media.