Research on Logistics Cargo Prediction Based on Machine Learning Methods

Abstract

Accurate prediction of cargo volume and reasonable arrangement of staff scheduling are crucial for improving the efficiency and punctuality of logistics transport. To address the problem of cargo volume prediction, this paper first conducts a visual analysis of the cargo volume data of sorting centres, and finds that the cargo volume of most sorting centres exhibits a rising trend of fluctuation over time. Then machine learning methods such as Random Forest and SVR are used to construct a prediction model by combining lagging features, trending features and periodic features. After model training and evaluation, the Random Forest model was finally selected to fit the prediction of the cargo volume of different sorting centres, and the rolling prediction strategy was used to ensure the timeliness and accuracy of the prediction results Improved accuracy of logistics cargo forecasting.