Hate and Offensive Speech Detection Using Machine Learning

The widespread occurrence of offensive and hateful texts on digital platforms, especially online media sites like Twitter, has sparked worries about user security and the welfare of the community. In order to recognize and reduce the impact of such harmful content, this article explains about a hateful and offensive language detection system using ML algorithms. Also the difficulties of identifying hateful and offensive content in a large volume of content produced by users are also discussed. This study uses a decision tree classifier as the main model and thoroughly compares and contrasts KNN, Naive Bayes (NB), and Random Forest (RF) classifying techniques to determine the best. Using labeled datasets from social media platforms for training and validation, the models are put through extensive testing to determine recall, accuracy, and precision. After analysis, we found that decision tree classifier is the best model. Due to its advanced features and classification methods, it produces the best accuracy rates, out of all other models. Referring this study, various better filtering solutions can be developed, which can help to make the Internet usage safer due to powerful ML algorithms that will help to prevent hate speech online.