A Comparative Supervised Machine Learning Framework for Sentiment Analysis of Amazon Product Reviews

Abstract

The paper explores the use of supervised machine learning algorithms, specifically Logistic Regression and Random Forest, for sentiment analysis of Amazon product reviews. With the rise of e-commerce, understanding customer sentiment from online reviews has become essential for businesses to improve their offerings and strategies. The study utilizes a dataset of Amazon product reviews, pre-processing the text through tokenization, stopword removal, and TF-IDF vectorization. Logistic Regression, a linear model, and Random Forest, an ensemble learning method, are employed to classify reviews into positive, neutral, or negative sentiments. The models' performance is evaluated using metrics like accuracy, precision, recall, and F1-score. Results indicate that while both models perform well, Random Forest outperforms Logistic Regression in terms of accuracy, though Logistic Regression offers greater interpretability and efficiency for large datasets. This research highlights the effectiveness of machine learning algorithms in extracting valuable insights from customer reviews, contributing to enhanced business decision-making and product development strategies.