A Machine Learning-Based Framework for Sentiment Analysis of YouTube Comments

This paper presents a YouTube Comment Sentiment Analysis system designed to automatically classify user comments into positive, negative, and neutral categories. The system leverages Natural Language Processing (NLP) and Machine Learning (ML) techniques to analyze large volumes of unstructured textual data. Using the YouTube Data API, comments are extracted and preprocessed through cleaning, tokenization, and vectorization. Multiple supervised learning models such as Naïve Bayes, Logistic Regression, and Support Vector Machine (SVM) are trained and evaluated. The system enables real-time sentiment prediction and visualization, helping in understanding audience feedback, engagement trends, and content performance