Classification of Online Toxic Comments Using Machine Learning

Abstract

The rapid growth of social media platforms, online discussion forums, and digital communication has significantly increased the volume of user-generated content. While these platforms promote communication and information sharing, they also face challenges due to toxic comments, including hate speech, cyberbullying, abusive language, threats, and offensive content. Such comments negatively impact users' mental well-being, discourage healthy discussions, and create unsafe online environments. Manual moderation of millions of online comments is time-consuming, expensive, and often inconsistent. This study proposes a machine learning-based online toxic comment classification system that automatically identifies and classifies toxic comments using Natural Language Processing (NLP) techniques. The textual data undergo preprocessing steps such as text cleaning, tokenization, stop-word removal, stemming, lemmatization, and TF-IDF vectorization before model training. Several supervised machine learning algorithms, including Logistic Regression, Naïve Bayes, Support Vector Machine (SVM), Decision Tree, Random Forest, and Extreme Gradient Boosting (XGBoost), are trained and evaluated using standard performance metrics such as accuracy, precision, recall, F1-score, and confusion matrix. Experimental results demonstrate that Logistic Regression and Support Vector Machine achieve high classification accuracy, while ensemble methods such as Random Forest provide robust performance. The proposed system offers an efficient and scalable solution for automated content moderation, helping social media platforms maintain respectful online communities and reduce the spread of harmful content.

Keywords