Dissertation

ABSTRACT

Jihadist terrorism represents a global threat for society and a challenge for scientists interested in understanding its complexity. This complexity continuously calls for developments in the ways in which terrorism is studied. Enhancing the empirical knowledge on the phenomenon can potentially contribute to develop concrete real-world applications and, ultimately, to the prevention of societal damages. Taking these aspects into consideration, this work presents a novel methodological framework that integrates network science and deep learning with the aim to shed light on jihadism, both at the explanatory and predictive levels. Specifically, this dissertation will compare and analyze the world's most active jihadist terrorist organizations (i.e. The Islamic State, the Taliban, Al Qaeda, Boko Haram, and Al Shabaab) to investigate their behavioral patterns and forecast their future actions. The analyses will employ hybrid techniques and pursue three linked aims. Firstly, using stochastic transition matrices to detect complex recurring behaviors in jihadist operational dynamics and presenting Normalized Transition Similarity, a novel coefficient of pairwise similarity in terms of strategic choices. Secondly, applying Hawkes point processes to investigate the presence of time-dependent structure in attack sequences. Thirdly, integrate complex meta-networks and deep learning to rank and forecast most probable future targets. This research seeks to reveal how hidden abstract connections between events may be exploited to unfold jihadist temporal mechanics and how memory-like processes (i.e., multiple non-random parallel and interconnected recurrent behaviors) can illuminate the way in which these groups act.