Challenges in Discriminating Profanity from Hate Speech

TL;DR

This paper investigates the challenge of distinguishing profanity from hate speech in social media, introducing a new dataset and classification methods that achieve up to 80% accuracy, highlighting the complexity of the task.

Contribution

It presents a new annotated dataset and applies supervised classification with various features and ensemble methods to differentiate profanity from hate speech.

Findings

Achieved 80% accuracy in 3-class classification

Surface n-grams are insufficient for deep understanding

Label variability due to subjective annotations

Abstract

In this study we approach the problem of distinguishing general profanity from hate speech in social media, something which has not been widely considered. Using a new dataset annotated specifically for this task, we employ supervised classification along with a set of features that includes n-grams, skip-grams and clustering-based word representations. We apply approaches based on single classifiers as well as more advanced ensemble classifiers and stacked generalization, achieving the best result of 80% accuracy for this 3-class classification task. Analysis of the results reveals that discriminating hate speech and profanity is not a simple task, which may require features that capture a deeper understanding of the text not always possible with surface n-grams. The variability of gold labels in the annotated data, due to differences in the subjective adjudications of the annotators, is also an issue. Other directions for future work are discussed.