Research on the Development of New Energy Vehicles Based on Multiple Machine Learning Models and the ARIMA Algorithm

This paper investigates the development of new energy vehicles in China using a combination of multi-model machine learning and the ARIMA algorithm. Initially, four indicators and six influencing factors representing the development of new energy-electric vehicles in China were identified. Through matrix forwarding and data standardization, normalized data was obtained. The entropy weight method was employed to calculate weights, resulting in development scores from 2011 to 2022. Subsequently, seven models including linear regression, polynomial regression, decision tree regression, random forest, gradient boosting, BP neural network, and support vector machine (SVR) were established. Different training and testing set proportions (70%, 75%, and 80^%) were used to determine the best-fitting effect. The model with the best fitting effect was identified, and the influencing factors were ranked based on their impact on the development of new energy-electric vehicles in China. Additionally, utilizing the four indicators, the LSTM model and ARIMA algorithm were employed to forecast, evaluate, and score the development indicators. Subsequently, based on 12 years of data from 2011 to 2022, the paper describes and predicts the development of new energy electric vehicles in China for the next 10 years.