Reseaech On Sales Forecast Model Of Electric Vehicle Based On TEI@I Method

The prediction method based on network search has independence and authenticity,but its practical application effect in electric vehicle sales forecasting needs further research and verification.At present,most scholars make national automobile sales forecasts based on macro factors such as population,policy and economy.A few scholars have also combined network search data to predict the sales of specific vehicle models.However,there are few studies on the sales forecast of electric vehicles.Based on the analysis of the electric vehicle sales forecasting model,this paper finds that the forecasting ability of the integrated model based on TEI@I methodology is better than the linear model.The main research contents and conclusions are as follows:1.Using Web of Science as the retrieval platform,the relevant literatures of the past 20 years were collected,and CiteSpace was used for cluster analysis of the literature to obtain citation clustering network map and time zone map.The clustering results show that the United States and China have more research on the automobile market forecasting and started earlier.Moreover,the current research hotspots mainly focus on the construction of vehicle sales forecasting models and the location of electric vehicle charging piles.At the same time,through literature review,summarizing the experience and methods of predecessors predicting the sales of electric vehicles,it is found that the TEI@I methodology proposed by Chinese scholars is widely used in prediction models.2.Starting from the consumer decision-making theory,the theoretical framework for the correlation between the influencing factors of macroeconomic data,Baidu index and word-of-mouth data and electric vehicle sales is established.By mining consumer network data,Granger causality test and grey correlation degree are used to quantitatively analyze the factors affecting the sales of electric vehicles.The research results of electric vehicle models show that the Baidu search index lags behind for three months is time-sensitive to the sales of electric vehicles.3.Guided by TEI@I methodology,using the idea of first decomposition and integration,first,using principal component regression analysis to fit the linear relationship in the prediction model,and then using BP neural network and support vector machine to fit the nonlinear relationship in the model,and finally,the integration.Taking the data of two car models as examples,it is found that the PCR-BP model and the PCR-SVM model have better prediction performance than the single model.In short,this paper constructs an integrated forecasting model for electric vehicle sales based on TEI@I methodology,and carries out an example analysis to verify that the integrated model is better.The actual forecasting effectiveness can provide an effective decision-making reference for similar product market forecasts.