<![CDATA[In recent years, the rapid growth of e-commerce has made customer reviews an important indicator of product quality, service performance, and customer satisfaction. Review scores play a crucial role in influencing purchasing decisions and evaluating overall shopping experiences on e-commerce platforms. This research aims to analyze the main factors influencing customer review scores by integrating logistics, transaction, and product-related variables using a machine learning approach. The dataset consists of various e-commerce transaction attributes, including delivery information, payment details, and product characteristics. A Random Forest classifier is employed to predict customer review scores and to identify dominant influencing factors through feature importance analysis. The results show that logistics-related factors, particularly delivery time, are the most influential variables affecting review scores, followed by payment value, freight value, and product price. This study also emphasizes the significance of understanding how models work and their real-world applications, offering useful guidance on enhancing logistics efficiency, ensuring clear transaction records, and maintaining high standards of product information. Product attributes such as description length, weight, and physical dimensions also contribute significantly to customer satisfaction. By combining predictive capability and interpretability, this research provides valuable insights for sellers and e-commerce platform managers to improve service quality, optimize logistics performance, and enhance customer satisfaction.]]>
Copyrights © 2026
