The study of alpha in GM(1,1) model

The purpose of this paper is to analyze a predicted error in using the GM(1 , 1) model based on the parameter alpha. The transfer function for the pred icted error with the parameter alpha in the GM(1, 1) model is presented. Th e algorithm of solving equations in calculus is used to analyze whether the ex is adaptive or not. The criterion of alpha is applied to describe the a daptive criterion of alpha. Finally, an example of the cage-net amounts of fish in the Peng-hu area is used to demonstrate the small prediction error due to the optimal alpha value. The result shows that the criterion for alp ha is applicable for minimizing the predicted error easily.