Monitoring Customer Satisfaction on Amazon through Sentiment Analysis using Supervised Machine Learning

In the era of digital commerce, platforms like Amazon produce a lot of customer reviews, which are helpful predictors of consumer satisfaction and buying trends.This study focuses on sentiment analysis of Amazon product reviews to track customer satisfaction using supervised machine learning techniques.Using algorithms like Decision Tree, Support Vector Machines (SVM), Random Forest, and more complex models like transformer-based architectures (BERT, RoBERTa), the system classifies reviews as either positive, negative, or neutral.Tokenization, stopword removal, and word embedding (TFIDF, Word2Vec, FastText) are a few preprocessing methods that increase model accuracy.Experimental results demonstrate that integrating contextual information from product headlines and titles significantly improves classification performance, with hybrid models achieving accuracies of over 70%.This technique makes it possible to track customer sentiment in real time, which yields useful information for marketing strategies, inventory control, and product development.By linking customer feedback with data-driven decision-making, the proposed framework assists e-commerce businesses in enhancing customer satisfaction and preserving a competitive edge.