Assessing violent group culpability and classifications using open-source intelligence

Because traditional secret intelligence can be difficult to obtain regarding subnational violent groups, open-source intelligence is considered a potentially valuable source of information. Open-source intelligence, obtained from news reports, blogs, and other publically available resources, is a rich, but noisy source of information about violent group activities. In an era during which collaborations across departments, agencies, or governments bring together experts with diverse training, resources, and objectives, developing methods for aggregating and objectively interpreting such sources of information is essential. In this paper, we explore the capability of combining open-source intelligence with computational methods to solve two problems relevant to understanding and responding to subnational group violence.