Automatic detection of cyber-recruitment by violent extremists

Abstract

Growing use of the Internet as a major means of communication has led to the formation of cyber-communities, which have become increasingly appealing to terrorist groups due to the unregulated nature of Internet communication. Online communities enable violent extremists to increase recruitment by allowing them to build personal relationships with a worldwide audience capable of accessing uncensored content. This article presents methods for identifying the recruitment activities of violent groups within extremist social media websites. Specifically, these methods apply known techniques within supervised learning and natural language processing to the untested task of automatically identifying forum posts intended to recruit new violent extremist members. We used data from the western jihadist website Ansar AlJihad Network, which was compiled by the University of Arizona's Dark Web Project. Multiple judges manually annotated a sample of these data, marking 192 randomly sampled posts as recruiting (YES) or non-recruiting (NO). We observed significant agreement between the judges' labels; Cohen's κ = (0.5, 0.9) at p = 0.01. We tested the feasibility of using naive Bayes models, logistic regression, classification trees, boosting, and support vector machines (SVM) to classify the forum posts. Evaluation with receiver operating characteristic (ROC) curves shows that our SVM classifier achieves an 89% area under the curve (AUC), a significant improvement over the 63% AUC performance achieved by our simplest naive Bayes model (Tukey's test at p = 0.05). To our knowledge, this is the first result reported on this task, and our analysis indicates that automatic detection of online terrorist recruitment is a feasible task. We also identify a number of important areas of future work including classifying non-English posts and measuring how recruitment posts and current events change membership numbers over time.

Value Predicate Analysis: A Language‐Based Tool for Diagnosing Behavioral Tendencies of Religious or Value‐Based Groups in Regions of Conflict

“…5 The majority of the literature employs sentiment analysis to detect or classify existing radical/hate speech or existing terrorist activities and recruitments in online environments rather than predicting future violent behaviors. See, for example, Azizan and Aziz (2017:691-98), Kocharekar and Jadhav (2017:285-87), Scrivens and Frank (2016:104-07), Ferrara et al(2016:22-39), Bermingham et al(2009:231-36), Johansson, Kaati, and Sahlgren (2016), Scanlon and Gerber (2014), and Ngoge (2016). For the use of sentiment polarity to detect hate speech, see Schmidt and Wiegand (2017).…”

“…(:22–39), Bermingham et al. (:231–36), Johansson, Kaati, and Sahlgren (), Scanlon and Gerber (), and Ngoge (). For the use of sentiment polarity to detect hate speech, see Schmidt and Wiegand ().…”

Community-Oriented Policing and Technological Innovations

“…Specifically, several research efforts have proposed methods for analyzing extremist Web forums for detecting users representing potential lone wolf terrorists and perpetrators of radical violence (Johansson et al 2013;Scanlon and Gerber 2014). Additionally, the use of social media by terrorist and extremist groups and the resulting social network perspectives have also been studied.…”

Identifying Right-Wing Extremism in German Twitter Profiles: A Classification Approach

“…They build automatic classifiers based on social network structure properties and keywords. While this work focused on detection of groups, Scanlon et al dealt with specific events of interaction, namely the recruitment of individuals [12] on specific extremist's websites. Their domain are Western Jihadists.…”