A New Method For the Estimation of Optimum Weight Coefficients of the Weight Changeable Combination Forecasts Bases on The Generalized Regression Neural Network
Geng Jian-jun · Shuxue de shijian yu renshi · 2011
In this paper,a new method for the estimation of optimum weight coefficients of changeable weight combination forecast model is proposed.The changeable weight coefficients of the in-sample are obtained through quadratic programming method of the optimization tool box of the Matlab.Then the weights of the out-sample are obtained via the Generalized Regression of Neural Network.A practical example shows the effect of the method is remarkable and the accuracy of forecasting is higher than other methods.So it expands the application fields of combination forecasting.