Comment Toxicity Tracker Using NLP with Emphasis on Machine Learning Algorithms

Abstract: The rise of online stages has driven to an uncommon volume of user-generated content, including comments on various forums, social media posts, and news articles. However, this abundance of user comments has also brought to light the issue of toxicity, where certain comments contain harmful, offensive, or inflammatory language that can negatively impact online discussions and communities. To address this issue, investigate centers on the advancement of a comment harmfulness location show utilizing Normal Dialect Handling (NLP) & Machine Learning. The proposed system leverages state-of-the-art NLP. By training these models on labelled datasets of toxic and non-toxic comments, the system learns to identify patterns and linguistic cues associated with toxic language. Key components of the system preprocessing steps to clean and tokenize the comments.Feature extraction using word embeddings or contextual embeddings, and model training using Machine Learning algorithms like neural networks, Random Forest Classifier model, etc . Evaluation measurements such as exactness, exactness, review, and F1-score are utilized to survey the execution of the prepared model