A Machine Learning-Based Approach to Combat Hate Speech on Social Media

Separating hate speech from specific instances of offensive language is one of the biggest obstacles in the study of hate speech in online life. Lexical disclosure methods are typically not very effective because they do not understand the two groups in any communication having particular words as hateful as a speech. In this proposal, large hate speech vocabularies were developed to collect tweets that contained terms of hate speech. A dimensionality reduction approach is also used to increase the degree’s correctness. This study divided tweets into three categories: those with hate speech, those with aggressive language, and those without either. We are getting ready to classify these distinct characterizations into many classes when we notice them. With 83% accuracy, the calculation of the machine learning-based approach using a dimensionality reduction technique is more accurate than other existing methods, such as 71.33% of Naïve Bayes and 80.56% of SVM.