30 Jun, 2021
Predicting Domestic Extremism and Targeted Violence: A Machine Learning Approach
University of Nebraska at Omaha. National Counterterrorism Innovation, Technology, and Education Center
From the document: "The report applies machine learning (ML) techniques to forecast where domestic extremist groups and active shooter incidents are most likely to occur in the United States. Identifying high-risk areas for these emerging threats is important for effective counterterrorism and conflict prevention, but complicated by the fact that policymakers often need to detect these threats at a stage when there might not be overt warning signs of violence. This report addresses this gap and directly supports Strategic Goals 1.1 and 1.2 in the June 2021 National Strategy for Countering Domestic Terrorism by providing 'data-driven guidance on how to recognize potential indicators of mobilization to domestic terrorism.' We develop and test two prototype machine learning models based on existing research about the causes of radicalization, ideologically-motivated violent extremism (IMVE), and targeted violence. First, we input information about these potential risk indicators as well as data about extremist actors and violent incidents to map patterns between 2017-2020. We then use this information to forecast which areas are at highest risk for extremism and active shooter incidents. As an extension, we also identify which areas in the maritime domain are most likely to experience active shooter incidents. The model's high level of accuracy suggests that these risk indicators are highly predictive of extremist operations and incidents. Overall, these models provide guidance for practitioners about where extremist actors and violent incidents are most likely to emerge moving forward."
-
URL
-
Authors
-
Publisher
-
Report NumberNCITE Reports, Projects, and Research No. 25; National Counterterrorism Innovation, Technology, and Education Center Reports, Projects, and Research No. 25
-
Date30 Jun, 2021
-
CopyrightOpen Access. Authors have granted unrestricted access.
-
Retrieved FromDigitalCommons@UNO: digitalcommons.unomaha.edu/
-
Formathtml
-
Media Typeapplication/pdf
-
Subjects
-
List
Details