The Establishment and Application of Combination Forecasting Model
|School||Xi'an University of Technology|
|Keywords||Combination model Grey Forecasting Grey Relational Analysis ARMA model Neural Network Model|
With the new theory of gray system theory , BP neural network technology used to predict the field of forecasting techniques has been a great development . Strong fusion and penetration of gray prediction model with a general prediction model , the combination of gray model with other models for analysis and forecasting can improve the prediction accuracy . Therefore, analysis and research on the combination of gray model with other models , combination forecasting model and forecast China's energy consumption . The main contents of this paper and the results are as follows : 1 . Combination of ARMA model based on gray prediction GM (1,1) model , triangular model and time series analysis predict touch -type TGMA ( 1,1 ) . In this model, gray forecasting GM (1,1) model fits the data sequence trend term , residual series of triangular and ARMA models capture system . Building a gray neural network combined forecasting model . The combination of improved BP neural network of the three gray prediction model ( gray prediction model of optimal prediction model GOM (1,1), unbiased gray prediction model , improved metabolism ) . Gray neural network combined forecasting model to make further improvements . Gray relational analysis of several factors to identify a greater impact on China's energy consumption , to them time series data as the input of BP neural network , such a comprehensive and integrated manner taking into account the impact factors of the energy consumption of the system , thereby increasing the prediction accuracy.