Hate Speech Detection in Social Networks using Machine Learning and Deep Learning Methods (The Science and Information)

Social media hate speech is becoming a bigger problem that affects constructive dialogue. Effectively identifying and tracking hate speech is essential. The study suggests a thorough method for identifying hate speech on Twitter that combines deep learning and conventional machine learning approaches. Using tweets classified as neutral, hate speech, or harmful language, the researchers created a solid dataset. They found that deep learning techniques, especially BiLSTM, are the most accurate for detecting hate speech after comparing CNN, LSTM, and BiLSTM models to conventional shallow learning approaches. BiLSTM models perform exceptionally well in terms of context sensitivity, semantic subtleties, and sequential patterns. Word embeddings, such as Word2Vec and GloVe, greatly improve model performance. The study emphasizes how effective deep learning methods are in thwarting hate speech and encouraging constructive online interactions.

https://thesai.org/Publications/ViewPaper?Volume=14&Issue=5&Code=IJACSA&SerialNo=42