An Intelligent XGBoost-Based Framework for Product Demand Forecasting Using Machine Learning

Abstract

Accurate demand forecasting plays a vital role in inventory planning, supply chain optimization, and production management by enabling organizations to anticipate future product requirements. This paper presents a machine learning-based demand forecasting framework that predicts future product demand using historical sales data and comparative predictive modeling. The proposed system performs data preprocessing through normalization, randomization, and dataset partitioning before training multiple forecasting models, including Random Forest, Gradient Boosting, Long Short-Term Memory (LSTM), and Extreme Gradient Boosting (XGBoost). The performance of each model is assessed using the coefficient of determination (R² score) and Root Mean Square Error (RMSE) to measure prediction accuracy and error magnitude. Experimental evaluation demonstrates that the XGBoost model achieves superior forecasting performance with an accuracy of 98%, outperforming Random Forest (84%), LSTM (87%), and Gradient Boosting (91%), while producing the lowest prediction error. A webbased application is developed to provide secure user authentication, dataset processing, model execution, comparative performance analysis, and real-time demand prediction for selected products. The proposed framework offers an efficient and scalable solution for demand forecasting, assisting organizations in improving inventory control, minimizing stock shortages, and supporting datadriven operational decision-making.