# Detection of Violent Extremists in Social Media

## TL;DR

This paper presents an automatic detection framework for identifying extremist users on social media, utilizing username, profile, and content features, demonstrated to be effective on Twitter data related to ISIS.

## Contribution

It introduces a novel detection scheme that combines multiple user-related features to identify online extremists more accurately than existing methods.

## Key findings

- Effective detection of ISIS-related extremists on Twitter
- Usernames tend to be similar among extremists and their like-minded
- The proposed framework outperforms baseline detection approaches

## Abstract

The ease of use of the Internet has enabled violent extremists such as the Islamic State of Iraq and Syria (ISIS) to easily reach large audience, build personal relationships and increase recruitment. Social media are primarily based on the reports they receive from their own users to mitigate the problem. Despite efforts of social media in suspending many accounts, this solution is not guaranteed to be effective, because not all extremists are caught this way, or they can simply return with another account or migrate to other social networks. In this paper, we design an automatic detection scheme that using as little as three groups of information related to usernames, profile, and textual content of users, determines whether or not a given username belongs to an extremist user. We first demonstrate that extremists are inclined to adopt usernames that are similar to the ones that their like-minded have adopted in the past. We then propose a detection framework that deploys features which are highly indicative of potential online extremism. Results on a real-world ISIS-related dataset from Twitter demonstrate the effectiveness of the methodology in identifying extremist users.