SENTIMENT CLASSIFICATION FOR ONLINE BOOK REVIEWS USING ENSEMBLE CLASSIFIERS

Abstract

Due to the rapid expansion of online shopping platforms, sentiment analysis of user-generated product reviews has become a crucial area of study in machine learning and natural language processing. Businesses can enhance customer satisfaction, product offerings, and marketing strategies by accurately gauging consumer sentiment. The goal of this paper is to improve the interpretability and dependability of sentiment classification systems by fusing ensemble approaches with potent feature extraction methods. To improve sentiment classification accuracy, this study examines the application of ensemble classifiers, such as Gradient Boosting Machine (GBM), Extreme Gradient Boosting (XGBoost), Adaptive Boosting (AdaBoost), and Categorical Boosting (CatBoost). By combining the advantages of several base models, ensemble approaches increase robustness and generalization when dealing with noisy, unstructured textual input. The study investigates feature extraction methods that capture semantic links and contextual subtleties in text, such as n-grams. To assess the model's performance, experiments were carried out on Amazon online book reviews datasets. The reported evaluation indicates that the use of ensemble classifiers produced good results in terms of accuracy, precision, recall, and F1 score. The reported results also indicate a sound foundation that will support scholarly paper as well as real-world e-commerce applications, helping companies use customer input to gain a competitive edge.