Deep learning framework for interpretable supply chain forecasting using SOM ANN and SHAP
· Source: nature.com
In short
This article proposes a deep learning framework integrating Self-Organizing Maps (SOMs), Principal Component Analysis (PCA), Artificial Neural Networks (ANNs), and SHAP explainability techniques to enhance supply chain forecasting accuracy and interpretability. Using large datasets, the model outperforms conventional machine learning methods in predicting shipping times and late delivery risks, achieving high accuracy and transparency that supports decision-making and resilience in supply chain management. The framework’s scalability and robustness are validated across multiple datasets, demonstrating its practical applicability and superiority over existing approaches.
Read Original Article — Read the full article on the publisher’s site.
In the issue of week 29, 2025: Agentic AI and Digital Twins Forge Resilient, Profitable Supply Chains of 2025
Tags: forecasting, logistics, optimization, classification, interpretability, deep learning