Data-Intensive Inventory Forecasting with Artificial Intelligence Models for Cross-Border E-Commerce Service Automation

Abstract

Building an adaptative, flexible, resilient, and reliable inventory management system provides a reliable supply of cross-border e-commerce commodities, enhances supply chain members with a flow of products, fulfills ever-changing customer requirements, and enables e-commerce service automation. This study uses an e-commerce company as a case study to collect intensive inventory data. The key process of the AI approach for an intensive data forecasting framework is constructed. The study shows that the AI model’s optimization process needs to be combined with the problems of specific companies and information for analysis and optimization. The study provides optimization suggestions and highlights the key processes of the AI-predicting inventory model. The XGBoost method demonstrates the best performance in terms of accuracy (RMSE = 46.64%) and reasonable computation time (9 min 13 s). This research can be generalized and used as a useful basis for further implementing algorithms in other e-commerce enterprises. In doing so, this study highlights the current trend of logistics 4.0 solutions via the adoption of robust data-intensive inventory forecasting with artificial intelligence models for cross-border e-commerce service automation. As expected, the research findings improve the alleviation of the bullwhip impact and sustainable supply chain development. E-commerce enterprises may provide a better plan for their inventory management so as to minimize excess inventory or stock-outs, and improve their sales strategies and promotional and marketing activities.

Keywords

• cross-border e-commerce
• supply chain management
• inventory forecasting
• data-intensive
• artificial intelligence
• Extreme Gradient Boosting