Comparative Study On Prediction Model Based On Vehicle Sales

The automobile industry is a pillar of the national economy, has veryimportant significance for the development of the national economy. In recentyears, with the development of automobile industry economy is booming, theautomotive industry expansion; increase the automobile market capacity; salespromotion in the application to remind and requirements of producers andmanagers more reasonable ways to deal with the changes in the market,develop a more reasonable production and sale plan to respond to marketdemand, then the littleprocessing and prediction of sales can not. Accuratelypredict the sales ofproducers and operators is particularly important, then themodel of how to use the method to predict the trend of better sales and futureis an important andworth further research.By analyzing historical data to establish the model of car sales, is thestarting point for the prediction of automobile sales. The goal is through theestablishment of a sales on vehicle flexible and effective prediction model,predictto achieve the sales. Prediction model and linear prediction modelusing the nonlinear prediction model, prediction method combined with thequantitativequalitative forecast method, and analyzed the test applicationhistorical sales data, and then establish the final prediction model forautomobile sales.Prediction model, by using the grey correlation analysis of influencefactors on automobile sales degree using the linear, and use GM (1,1) modelof the BP neural network prediction model, using classical modeling of carsales to car sales, and the use of BP network to the GM (1,1) model ismodified, improved prediction precision. Through empirical analysis based onthe comparison of the characteristics of the model, the application conditionsand application model of the respective advantages of each model andcombination, form a linear model, RBF neural network and grey theory ofcombination forecasting model.