Classification of Online Toxic Comments Using Machine Learning Algorithms

Abstract-The increasing prevalence of toxic comments on social media necessitates efficient automated systems for content moderation. This paper presents a machine learning-based approach to classifying toxic comments, aiming to detect harmful content such as hate speech, threats, and offensive language. We evaluate various supervised learning algorithms, including logistic regression, support vector machines (SVM), random forests, and advanced deep learning models such as recurrent neural networks (RNNs) and transformer-based models like BERT. Text preprocessing techniques like tokenization and feature extraction using TF-IDF and word embeddings are applied to optimize model performance. The models are trained on large labeled datasets and evaluated using accuracy, precision, recall, and F1-score. Our results show that deep learning models, particularly transformer-based architectures, achieve superior performance in identifying toxic comments, highlighting their effectiveness in supporting content moderation on social media platforms.