Optimization of used price assessment model for new energy vehicles using machine learning

The booming development of the new energy vehicle industry has led to the development of the new energy vehicle used market. The study introduces machine learning into the price evaluation of new energy used cars to alleviate the cost of new energy used transactions and improve the transaction efficiency. After collecting and pre-processing the transaction data of new energy automobile second-hand market, the valuation index system is established by selecting new energy second-hand vehicle valuation candidate indicators. The mean absolute error (MAE), root mean square error (RMSE), R-squared value (R2) and mean absolute percentage error (MAPE) are used as the indicators to measure the valuation performance. Other machine learning algorithms are fused into the Stacking algorithm to construct a new energy used car valuation model based on Stacking fusion. Empirical analysis is conducted to test the effectiveness of the Stacking valuation model on new energy price assessment. Among the first 100 samples to be valued, the maximum price difference between the appraisal price of the Stacking fusion valuation model in this paper and the actual results of the samples is 45,800 yuan. The percentage of absolute error in the appraisal of the traditional market approach for the cases to be appraised reaches 14.70%, which is much higher than that of the Stacking fusion valuation model. The valuation goodness of fit of the Stacking fusion valuation model and the SVM model in this paper are 0.989 and 0.875, respectively. The valuation price error of the traditional market comparison method fluctuates greatly between −8~80,000 yuan, and the valuation error of the Stacking fusion valuation model in this paper is concentrated in the range of [−2~2]. The Stacking fusion valuation model in this paper has obvious advantages in the valuation of new energy used cars.