Product order-demand prediction model based on random forest

Abstract

This study aims to develop a decision support model based on product order data analysis and demand forecasting. By analyzing the shipment data of a large manufacturing enterprise from September 2015 to December 2018, we establish an accurate prediction model for the demand in the next three months of a large manufacturing enterprise. Quarterly and monthly variables capture trends and seasonal variation by adjusting hyperparameters and cross-validation using a random forest algorithm. The results show that the mean absolute error (MAE) on the test set is 8.965, the root mean square error (RMSE) is 11.369, the relative mean absolute error (MAPE) is 8.256%, and the coefficient of determination (R²) is 0.826. These indicators confirm that the model can accurately predict the target variable, with little difference from the true value, and show good predictive power and fit. The monthly model has high accuracy and stability and can effectively support production and supply chain planning to meet future needs. This study confirms the potential of product order data analysis and demand prediction models to improve the efficiency and competitiveness of enterprises and provides a valuable reference for the research and practice in related fields.