INTELLIGENT PREDICTION OF CROSS-BORDER E-COMMERCE CUSTOMER SATISFACTION USING DEEP LEARNING EMBEDDINGS

Abstract

The rapid growth of cross-border e-commerce has intensified the need for intelligent systems capable of understanding and predicting customer satisfaction across diverse geographical, cultural, and logistical contexts. Customer satisfaction is influenced not only by textual feedback but also by multiple factors such as delivery time, pricing, product category, and country-specific expectations. This project presents an AI-driven web-based system that predicts customer satisfaction by leveraging deep learning embeddings derived from customer reviews while integrating contextual e-commerce attributes. The system utilizes a transformer-based deep learning model to extract semantic representations from customer reviews and combines them with structured transactional data to generate accurate satisfaction predictions. A role-based web architecture is implemented to manage users, training workflows, and analytical insights. Experimental results demonstrate that the proposed system effectively captures customer sentiment and provides actionable insights for administrators and stakeholders, thereby enhancing decision-making in global e-commerce platforms.