Understanding the psychology of online extremism promoters

Leveraging socio-psychological features for prediction of extremist users on Twitter

Mentor(s): Dr. Swati Agarwal, Domain(s): NLP , Social Network Analysis

PCA Plots for Train, Val and Test with Threshold = 0.8 for separating Positive and Negative extremist tweets.

Built a system for analyzing and identifying extremism promotion in tweets using ML, and detecting relevant features to prevent the same in a semi-supervised fashion. The tweets were collected using tools like Twitter API (Twython), and the pre-processing and feature-engineering was performed using GrammarBot, LiWC and TextRazor APIs.

Paper coming soon!