We can identify the benefits and drawbacks of the state-of-the-art methods that are currently in use by taking into account a variety of factors to obtain a thorough understanding. The results of our study show that deep learning-based approaches do not perform as well in detecting hate speech as traditional learning-based approaches do. This work highlights the importance of conventional machine learning techniques in successfully handling hate speech detection tasks, even though it acknowledges the promise shown by deep learning approaches.
https://ieeecai.org/2024/wp-content/pdfs/540900a328/540900a328.pdf