World Terrorism Attack Analysis: A Data Science Approach

Aaron A Gauthier
2020 Zenodo  
There are over 200 different definitions of terrorism. Violence is the hallmark of terrorism. A qualifying terrorist incident is commonly associated with a political motive. Understanding the attributes involved in terror attacks is a profoundly complex task. Psychologists would agree that terrorists are not highly likely to volunteer as experiment test subjects to be psychoanalyzed. Examining their activities from afar can be quite misleading as assumptions have to be made. The psychology
more » ... d terrorism due to all these complexities proves to be challenging to say the least. Understanding the patterns and behaviors behind the attacks are important to unlocking the prediction power of attacks in the future and their mitigation. The goal of my research will be to build a recommendation engine to augment human decision making. The goal of this project will be to demonstrate the use of Classical ML as a baseline technique to accurately predict which group(s) could be responsible for an attack based on certain attributes within the data. This may lead to additional research utilizing Deep Learning (DL) to create a recommendation engine that will inform and shape decisions makers thinking. The ML component will primarily rely on Python, Pandas, and SciKitLearn.
doi:10.5281/zenodo.3813858 fatcat:sxxdlyu3afbjdc55iav2av253q