This research advances hate speech detection by exploring multi-label classification approaches instead of traditional binary methods, better reflecting the nuanced nature of harmful content.

• Focuses on English-language hate speech requiring differentiated classification
• Surveys existing machine learning models and datasets for multi-label hate speech detection
• Provides a systematic framework for practitioners in content moderation and law enforcement
• Addresses critical security needs for protecting online communities and individuals

The multi-label approach significantly improves content moderation systems by enabling more nuanced responses to different types of harmful content, supporting more effective security measures across digital platforms.

A Survey of Machine Learning Models and Datasets for the Multi-label Classification of Textual Hate Speech in English