Empirical Analysis Of New Energy Vehicles Based On Statistical Prediction Model

Since the 21st century,automobiles have become visible everywhere in China,and gradually become an indispensable part of the lives of most residents.However,it is precisely because of the rapid economic expansion that it has also had a certain impact on ecology of China.In this way,green development has become very necessary.And not only is it a private appeal,but the development of new energy automobiles has been strongly supported by the government in recent years.The new energy automotive industry has been booming in China for more than a decade.In recent years,the production and sales level of China’s clean energy vehicle industry has also achieved a qualitative leap.In November 2020,China’s State Council issued the new energy vehicle industry development plan 2021-2035,It is clearly stipulated that the sales volume of new cars produced by China’s clean energy vehicles will exceed about 20% of the total sales volume of new cars in China by 2025.Therefore,the sales forecast for the production of clean energy vehicles industry will be more necessary,and reasonable expectation can also help China’s new energy vehicle manufacturing company to carry out more reasonable regulation and control according to the enterprise strategy.This paper adopts the linear mixed model,selects the data from 2016 to 2020,takes the sales volume of new energy vehicles(y)as the response variable,and takes the production volume of new energy vehicles(PONEV),automobile production volume(CD),automobile sales volume(CS),gasoline price(OP),consumer confidence index(CCI),number of public charging piles(NOPCP),consumer price index(CPI)Per capita disposable income(PCDI)is the original independent variable.Three prediction models are constructed and recorded as models 1,2 and 3respectively.The independent variables of the model come from the common factors screened by factor analysis,the significant variables screened based on the linear full model and the significant variables screened by the new variable selection method.Finally,according to the results of the model,the effect of model 1 is worse than that of models 2 and 3.Through the comprehensive selection of AIC,BIC,loglik,MAPE and,it is concluded that the effects of models 2 and 3 are equal and have better prediction effects.It can be explained that the new variable selection method used in this paper is applied to the linear hybrid model,which has better prediction effect on the sales volume of new energy vehicles,Then,when the relevant independent variables are known,the sales volume of new energy vehicles can be reasonably predicted through model 3.