New Energy used Car Residual Value Prediction Method Based on GBDT Model + Computer Simulation Technology

Accurate prediction of the residual value of used new energy vehicles is crucial for the healthy development of the used car market. This paper first builds calculation indicators based on big data crawling technology. Subsequently, it simplifies the calculation indicators based on the distribution of data and the relationships between indicators. Then, leveraging the advantages of GBDT models, a model is trained using the simplified indicators for predicting the residual value of used new energy vehicles. Finally, the trained model, along with initialized parameters, is applied to predict the test dataset, achieving an average prediction accuracy of 96.6%, effectively meeting industry development needs. This paper provides the used new energy vehicle industry with a high-accuracy, robust model, and theoretical approach for calculating residual values, promoting the industry's high-quality development.