An integrated deep-learning and multi-level framework for understanding the behavior of terrorist groups

Dong Jiang; Jiajie Wu; Fangyu Ding; Tobias Ide; Jürgen Scheffran; David Helman; Shize Zhang; Yushu Qian; Jingying Fu; Shuai Chen; Xiaolan Xie; Tian Ma; Mengmeng Hao; Quansheng Ge

doi:10.1016/j.heliyon.2023.e18895

An integrated deep-learning and multi-level framework for understanding the behavior of terrorist groups

Heliyon. 2023 Aug 6;9(8):e18895. doi: 10.1016/j.heliyon.2023.e18895. eCollection 2023 Aug.

Authors

Dong Jiang^{1

2}, Jiajie Wu^{1

2}, Fangyu Ding^{1

2}, Tobias Ide³, Jürgen Scheffran⁴, David Helman^{5

6}, Shize Zhang⁷, Yushu Qian¹, Jingying Fu^{1

2}, Shuai Chen^{1

2}, Xiaolan Xie^{1

2}, Tian Ma^{1

2}, Mengmeng Hao^{1

2}, Quansheng Ge¹

Affiliations

¹ Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, 100101, China.
² College of Resources and Environment, University of Chinese Academy of Sciences, Beijing, 100049, China.
³ Centre for Biosecurity and One Health, Harry Butler Institute, Murdoch University, Murdoch, 6150, Perth, Australia.
⁴ Institute of Geography, Center for Earth System Research and Sustainability, University of Hamburg, Hamburg, 20144, Germany.
⁵ Institute of Environmental Sciences, Department of Soil and Water Sciences, The Robert H. Smith Faculty of Agriculture, Food & Environment, The Hebrew University of Jerusalem, Rehovot, 7610001, Israel.
⁶ Advanced School for Environmental Studies, The Hebrew University of Jerusalem, Jerusalem, Israel.
⁷ Computer Network Information Center, Chinese Academy of Sciences, Beijing, 100190, China.

Abstract

Human security is threatened by terrorism in the 21st century. A rapidly growing field of study aims to understand terrorist attack patterns for counter-terrorism policies. Existing research aimed at predicting terrorism from a single perspective, typically employing only background contextual information or past attacks of terrorist groups, has reached its limits. Here, we propose an integrated deep-learning framework that incorporates the background context of past attacked locations, social networks, and past actions of individual terrorist groups to discover the behavior patterns of terrorist groups. The results show that our framework outperforms the conventional base model at different spatio-temporal resolutions. Further, our model can project future targets of active terrorist groups to identify high-risk areas and offer other attack-related information in sequence for a specific terrorist group. Our findings highlight that the combination of a deep-learning approach and multi-scalar data can provide groundbreaking insights into terrorism and other organized violent crimes.

Keywords: Deep learning; Terrorism; Terrorist group; Terrorist network.