The Combined Forecasting Based On ARIMA Model

The economic prediction makes policies services for economy, the goal is improve the science level of economy management,and reduce the blindness of economy decision,it is important to hold the economic development and understanding the change trends of the future.Firstly, this paper studies the theories and methods about short-term force asting,After synthetically introducing the current situation of the common-used theory, and explains in detail the research of the ARIMA forecasting theory and its applied situation, then proceeding from request for forecasting, The paper proposes time series model ,fray prediction and BP network predicts the project.. And introduce the wavelet analysis Mallat algorithm and weight structure algorithm, This paper give neural networks , ARIMA model and gray forecast model of design-steps, gives examples of ARIMA-based authentication, the structure being chapters and sections combination model is built laying down basis.Secondly, this paper introduces combination model and sets up tentatively an combined forecasting and often offers the data theory method appears and design-steps by the principle of minimum square error,this combinational model combines Gray model and ARIMA theory. the methods are applied to the GDP of china .Finally, a combinatorial forecast method based wavelet analysis using Artificial Network and autoregressive integrated Moving Average (ARIMA) models is presented, the economic data is decomposed by wavelet transform and reconstructed respectively, the approximate sequence and the detail sequence in different frequency are obtained , the approximate prediction adopts by ANN , after the property of detail sequence is detected ,the data are forecasted by the suited BP and ARIMA Models respectively, finally the forecasted results of the sub-sequence are reconstructed propose as the final forecast results.