Deep Decision Analytics for Customer Review Understanding and Predictive Opinion Intelligence

Abstract

Customer reviews play a crucial role in today's e-commerce ecosystem by providing valuable information about customer satisfaction, product performance, and the overall shopping experience. As online retail continues to expand, an enormous amount of textual feedback is generated every day. Conventional evaluation approaches, such as average product ratings and the total number of reviews, offer only a broad assessment and are unable to fully capture the detailed opinions and emotions expressed within review texts. Consequently, a significant amount of useful customer feedback remains unexplored. In addition, manually examining large volumes of review data is both labor-intensive and impractical, creating a strong demand for intelligent automated analysis techniques. To overcome these limitations, this study presents a machine learning-driven framework for customer review analysis that predicts both product recommendation decisions and product ratings. The proposed framework includes data preprocessing, exploratory data analysis, and textual feature extraction using the Term Frequency–Inverse Document Frequency (TFIDF) technique. Several machine learning algorithms are employed, including Restricted Boltzmann Machine (RBM) integrated with Logistic Regression (LR) for classification, RBM combined with Ridge Regressor (RR) for regression, Gradient Boosting (GB), Extreme Gradient Boosting (XGB), and a Multi-Task Neural Network with Extra Trees (MTNN-ET), where these models operate using the Classification and Regression Tree (CART) methodology. The experimental findings indicate that the proposed MTNN-ET-CART model outperforms the existing approaches, achieving a classification accuracy of 0.9640 and a regression R² score of 1.0000. The framework efficiently handles large-scale customer review datasets while delivering accurate prediction results, thereby supporting informed decision-making and enhancing the overall customer experience across ecommerce platforms.