New energy vehicle quantity prediction model based on improved grey BP neural network
In order to predict the number of new energy vehicles scientifically,the initial value of the traditional grey model GM(1,1)was improved on the basis of exploring the adaptability of the grey prediction model.By using correlation analysis,12 key indicators that can accurately describe the internal characteristics of the development of new energy vehicles were selected to correct the residual error of the improved grey model,and a prediction model of new energy vehicle ownership combined with improved grey BP neural network was established.The results showed that the effective combination of the improved grey theory with undetermined initial value and the neural network had a very high precision of the training value of the new energy vehicle ownership,and accurately predicted the number of new energy vehicles by combining the factors,such as social development level,road conditions and the number of car drivers,etc.,in the previous year.The simulation predicted that the ownership of new energy vehicle across the country in 2023 would be 20.262 4 million units,with an absolute error of 147 600 units and a relative error of only 0.72%.
neural networknew energy vehiclescorrelation analysisimproved grey model
NETL
NSTL
万方数据
