The Research on Combination Forecasting Model of the Automobile Sales Forecasting System

Automobile sells system plays an important role in automobile sales area, through the whole produce and management. Some forecast models have had unilateralism in some side nowadays, such as ARMA model. For example, the data of non-linearity has some error by ARMA model. This paper, assembles curve –regression model, Time Series Decomposition Model and RBF neural networks according to the weight distribution. Putting the Data Mining, math statistics and neural networks technique into automobile sales forecast system, It can improve the problem of unilateralism, The Combination forecasting model improves the veracity and utility range in automobile sales forecast. This paper, can also be used in which some other economic data that take on the obvious time character and trend in car-making and selling.