China’s GDP Forecasting Model Based On Time Series Analysis

GDP(Gross Domestic Product) is a key index of national economic accounting, often regarded as the best index of a measure of economic condition of a country(or region). We can determine economic ups and downs of a country(or region) from the change of GDP.Time series analysis forecasting method reveals the law of the phenomenon changing with time through historical data sequences. The law extends to the future, so as to predict the future trend of the phenomenon. It has played an important role in the economic field. Time series analysis method can be applied to forecast the GDP of China, and using time series model can accurately predict the future trend of the GDP of China. This provides theoretical guidance for the effective regulation of national macro economy and policy making.In this paper, the theory is based on time series analysis. The paper uses the 57 quarter data of GDP aggregate-value of China from the first quarter of 2000 to the first quarter of 2014 released by the National Bureau of statistics, drawing support from EViews 6.0 software, SAS 9.1.3 software and Matlab 2011 a software, to simulate and analyze data so as to establish multiple seasonal ARIMA model, composite model and dummy variable regression model. The paper innovates bran-new Gauss function and sum of Sine function composite model and opens up the new idea of establishing nonlinear regression model of multiple dummy variables. Then forecasting of China’s GDP of the latter three quarters in 2014 can be given through the econometric models. After analysis on the accuracy of the three kinds of models, the model with relatively minimal prediction error is choosen as relatively optimal prediction model. It is of great theoretical and practical effect.