New Energy Vehicle Sales Forecast Based on GA-SFA-SVR

Abstract

With the proposal of a carbon peak carbon neutralization target, new energy vehicles instead of traditional fuel vehicles have become the trend. Accurate prediction of new energy vehicle sales is significant to market policy formulation, company strategy adjustment and carbon reduction target realization. Because of the characteristics of new energy vehicles, such as short development time, less data and great influence by policy, this paper first uses the grey analysis method to calculate the correlation coefficient of each index, and then uses the slow feature algorithm to extract the factors with the slowest change, and uses the SVR to predict. In order to verify the validity of the model, Bayesian regression and linear regression models are used for comparison. The conclusion shows that the prediction error of this method is kept below 2 %, and the prediction accuracy is the best.