Hate Speech Detection using Deep Learning and Text Analysis

Hate speech is any form of speech, gesture, written or physical expression that threatens a person or a group based on their race, ethnicity, religion, gender, sexual orientation, nationality, disability, or any other characteristic that is protected by law. Hate speech can take many forms, ranging from verbal harassment to physical violence. Hate speech detection has become an important task in NLP due to the growing frequency of hate speech on online forums and social media. The proposed research work aims to improve hate speech detection by doing modification in standard i.e., Modified bi-LS TM model vs RCNN. The study examines how well the modified model performs on tasks involving the classification of hate speech when compared to a conventional LS TM model. The improved bi-LS TM model is intended to capture the context and relationships more accurately between the words in hate speech utterances.The study uses a publicly accessible dataset of tweets containing hate speech and tweets without any hate speech. The proposed model is trained and tested with the help of various performance metrics such as F1-score, accuracy and precision, recall. The research outcomes show that the proposed model outperforms the standard IS TM model in detecting hate speech.