This project aims to develop a machine learning model to categorize the sentiment of YouTube comments, establishing an efficient method for real-time processing of large volumes of text data. It is made by effectively choosing and comparing different machine learning models we learned in class using relevant performance metrics.
I designed and implemented a machine learning model using Python to analyze user sentiment in YouTube comments, enabling scalable insight into audience feedback, trends, and online behavior across 18K+ real-world data points. I also did Exploratory Data Analysis (EDA), handling missing data, class imbalance analysis, and feature engineering (count vectorization, language categorization, comment length, and t-SNE visualization) to ensure the best results. Finally, I compared and tuned the multiple models using precision, recall, F1-score, confusion matrices, and accuracy, analyzing limitations and proposing future improvements based on experimental results.