Searching for the Right Stuff: Identifying Radical Users in Online Discussion Forums

• In Event: Examining the Behaviors and Characteristics of Terrorists and Extremists On and Off-line

Abstract

One of the most significant policy issues for law enforcement agencies is to identify radical users online. Yet in the last ten years alone, it is estimated that the number of individuals with access to the Internet has increased three-fold, thus leading to a constant flood of data. These conditions have necessitated guided data filtering methods, those that can side-step the laborious manual methods that have been classically used to identify relevant information. In this study, we continue to explore this underdeveloped area of research by applying our algorithm, Sentiment-based Identification of Radical Authors (SIRA), to a sub-forum of the most popular white supremacy discussion forum: Stormfront. While SIRA can quantify an author’s web-forum behavior that may be deemed as “extreme” by measuring their average sentiment score, volume, severity and duration of negative posts using a keyword-based sentiment analysis tool, we extended this algorithm in two meaningful ways. First, users’ online sentiment was measured around specific topics – i.e., adversaries of the extreme-right (keywords associated with Blacks, Jews, and LGBTQs). Second, sentiment was assessed over time, wherein trajectory groups were constructed using a semi-parametric group-based modeling approach. Results highlight the benefits and real-world implications of applying temporal SIRA to big data.

Authors

• Ryan Scrivens, Simon Frasier University

• Richard Frank, Simon Frasier University

• Garth Davies, Simon Fraser University