Hate Speech Detection Using the GPT-2 and Natural Language Processing

Hate speech on social media has increased as people use it to communicate and express their thoughts. Automatic hate speech detection has been studied in several languages. This study investigates the conclusion of the performance evaluation of the machine learning model GPT-2 (Generative Pre-trained Transformation 2) for natural language processing. The proposed model is developed and evaluated using a hate speech dataset. We attained an accuracy of about 97.12 percent, higher than the existing models called biLSTM, CNN, CNN-LSTM, and BERT.