Sentimental Analysis on Product Review in E-Commerce Platform using Machine Learning and Deep Learning

Abstract: This paper focuses on developing a comprehensive sentiment analysis system for customer reviews, combining traditional machine learning and advanced deep learning techniques. The system classifies reviews into positive, negative, or neutral categories through robust text preprocessing, feature extraction, and model training. Traditional classifiers like Random Forest, Naive Bayes, Logistic Regression, and SVM are utilized alongside advanced NLP models such as VADER for quick analysis of short reviews and BERT for an in- depth understanding of longer, context-rich reviews. The system employs ensemble methods to enhance accuracy and consistency in sentiment classification. It evaluates performance through metrics such as accuracy, precision, recall, and F1-score to ensure reliability and scalability. A user- friendly Flask-based web application enables seamless dataset uploads, real-time analysis, sentiment visualization, and downloadable results. The project aims to provide an efficient and accurate sentiment analysis solution adaptable to diverse e-commerce platforms and customer feedback scenarios

Keywords: Sentiment Analysis; NLP; VADER; BERT; Customer Reviews; TF-IDF.