Explainable AI in Demand Forecasting Machine Learning Models

This research article discusses the integration of Explainable Artificial Intelligence (XAI) in demand forecasting using machine learning (ML) models. It outlines the significance of demand forecasting in supply chain management and highlights the challenges posed by the complexity of ML models, which can hinder transparency and trust among stakeholders. The article presents various XAI methods, including SHAP, LIME, permutation importance, partial dependence plots, and counterfactual explanations, detailing their roles in enhancing interpretability and user trust in ML predictions. A comparative analysis of different ML models, such as Linear Regression, Random Forest, XGBoost, and LSTM, is provided to illustrate how these explainability techniques can reveal the internal logic of model decision-making processes. The findings indicate that incorporating XAI methods not only improves the accuracy of demand forecasts but also facilitates better communication of insights to business stakeholders, ultimately supporting strategic decision-making in retail and supply chain contexts.