DEMAND FORECASTING USING ML

Abstract

Accurate demand forecasting is essential for effective supply chain management, inventory control, and business decision-making. Traditional forecasting methods often fail to capture complex patterns in demand data, leading to inaccurate predictions. This project proposes a machine learning-based approach for demand forecasting by comparing multiple algorithms, including Random Forest, Gradient Boosting, and Long Short-Term Memory (LSTM), along with an advanced model, XGBoost. The dataset is preprocessed using techniques such as normalization, shuffling, and train-test splitting. Each model is trained using 80% of the dataset and evaluated on the remaining 20% using performance metrics such as R² score (accuracy) and Root Mean Square Error (RMSE). Among all models, XGBoost demonstrates superior performance with the highest accuracy and lowest error rate. The system is implemented using Python and deployed through a web interface, allowing users to select products and forecast demand in real time. The results show that machine learning models, especially XGBoost, provide highly accurate demand predictions, making them suitable for real-world business applications.