Data-Driven Network Freight Price Forecasting: Based on CatBoost and Isolation Forest

The development of big data and internet technologies has given rise to network freight platforms. The network freight platform has improved the efficiency of freight dispatching, and the pricing accuracy of the platform is a key factor in its successful operation. At present, the main method of freight pricing is manual pricing by dispatchers, which has the problem of inaccurate pricing caused by subjective human factors. This paper proposes a machine-learning pricing method based on CatBoost and Isolation Forest. Firstly, the Isolation Forest anomaly detection method is used to delete the abnormal data in the training data to ensure the quality of the training data set. The CatBoost pricing model is then trained using the processed training dataset, which is capable of capturing price patterns from historical data. Compared with the comparison model, the pricing accuracy has been improved. The method proposed in this paper provides a new reference tool for the logistics industry.