Investigating Deep Learning Approaches for Hate Speech Detection in Social Media : Portuguese-BR tweets
Leila Weitzel, Thalessa Hungerbühler Daroz, Luan Pereira Cunha, Rafael Von Helde, Lucas Mendonça de Morais · 2023
Social media constitutes a very open social network space, which lacks of barriers to access. It has become a way for some people to vent anger, fear, happiness, hate, hope, love and sadness without feeling self-conscious. Hate speech and offensive language published and diffused via online environments have the potential to cause harm and suffering to individuals and lead to social disorder beyond virtual space. Over the last decade, machine learning and natural language processing approaches have been developed to detect harmful user content on social media. This work aims to analyze and detect hate speech, on Portuguese language, during the2018 Brazilian presidential electioneering period. In order to achieve this goal, we experiment different configurations of Deep Learning networks, hyperparameter and features. As a result, the paper presents best settings of parameters for practical applications of hate speech mining. We also presented a Portuguese hate lexicon and a labeled dataset. All resources are available on our GitHub.