Study on Freight Volume Prediction of Routes Based on Random Forest Model

Abstract

With the rise of e-commerce and the ever-changing consumer demand, logistic network sorting centers face increasingly complex and significant challenges in cargo volume forecasting. To predict the cargo volume for various routes, this paper selects the Random Forest model for forecasting. Feature data were extracted, and the dataset was divided into a training set and a test set. The model was evaluated using the test set. Ultimately, a high coefficient of determination and low mean squared error were achieved, demonstrating the accuracy and reliability of the Random Forest model's predictions. The transportation scenario was extended to changes in transportation routes, focusing on the impact of adding and canceling transportation routes on cargo volume forecasting. By establishing model features, conducting feature synthesis, and calculating impacts, the adjusted cargo volume for each sorting center was obtained, and an analysis of changes in transportation routes was conducted. The influence of added and canceled routes was integrated into the cargo volume forecast, resulting in the final cargo volume prediction. By distributing the adjusted volume evenly on a daily basis, the accuracy and reliability of the forecast results were ensured. Finally, a visual representation of the forecast results was provided, intuitively showcasing the model's effectiveness.