Text Mining to Detect Insults in Online Discussion

Abstract

The anonymity of online communication makes it particularly prone to hostility. Given text from online discussions [9], we train a variety of supervised learning algorithms to classify insults directed at other forum members. After meticulous text processing to generate features, we fit naive Bayes and regularized logistic regression to establish a baseline… (More)

Topics

10 Figures and Tables

Slides referencing similar topics