Enhancing hate speech detection with user characteristics

Social media provide users with a powerful platform to share their ideas. Using one’s right to expression to incite hate toward a particular group of people is inappropriate. However, hate speech is pervasive in our society. Spreading hate through online social networks like Facebook, Twitter, Tiktok, and Instagram is commonplace in today’s milieu. One such case is the unprecedented COVID-19 pandemic, which engendered anti-Asian hate. In the current literature, there is limited study on using user features in conjunction with textual features to detect hate speech. In this paper, we propose to combine tweet textual features with a variety of user features to improve the state-of-the-art hate speech detection techniques. The user feature we propose consists of demographic, behavioral-based, network-based, emotions, personality, readability, and writing style. To test our approach, we used four different English datasets gathered from Twitter and available in the public domain. Our results show that combining tweet textual features with the proposed user features improves hate speech detection up to +0.32 in F1 score and beats previously proposed approaches that use a limited number of user features. The analysis of the most important features confirms that hateful tweets or their authors express more negative emotions and use more swear words.