Predicting sales and cross-border e-commerce supply chain management using artificial neural networks and the...

Area of Science:

• Supply Chain Management
• Machine Learning
• Internet of Things

Background:

• E-commerce relies on efficient distribution and demand forecasting for cost reduction.
• Cross-border e-commerce presents unique supply chain challenges.
• Integrating advanced technologies is crucial for optimizing e-commerce operations.

Purpose of the Study:

• To improve supply chain efficiency in cross-border e-commerce.
• To develop a robust system for order prediction and resource allocation.
• To leverage machine learning and IoT for enhanced e-commerce logistics.

Main Methods:

• A two-stage approach: order prediction followed by resource allocation.
• Utilizing a weighted mixture of neural networks for sales order forecasting.
• Employing the Capuchin Search Algorithm (CapSA) to optimize model weights and ensemble performance.

Main Results:

• Achieved a Root Mean Squared Error (RMSE) of 2.27 in sales prediction.
• Demonstrated a minimum decrease of 2.4 in RMSE compared to existing methods.
• Reduced Mean Absolute Percentage Error (MAPE) by 14.67% against comparative approaches.

Conclusions:

• The proposed method effectively predicts sales orders and optimizes resource allocation in cross-border e-commerce.
• The integration of neural networks and CapSA significantly enhances prediction accuracy.
• The approach offers a substantial improvement in supply chain efficiency and cost reduction.