A Novel Feature Extraction Classifier for Hateful, Offense, and Neutral Content on X

With the rapid growth of online social networks, content censorship remains controversial, dividing people into two groups, one supporting hateful content and the other supporting neutral content. This paper addresses the problem of classifying a tweet as hateful, offensive, or neutral content, which uses Term Frequency Inverse Document Frequencies (TFIDFs) for feature extraction. It uses the X dataset to train the proposed classifier model, and the results show that Gaussian Naive Bayes is the best-performing model after hyperparameter tuning of TFIDF features.