Simulation Error Characteristics of Grey Model GM(1,1) under Translation Transformation

Abstract

To reveal the change law of the simulation error of grey model GM(1,1) with translation transformation applying on the original sequence, in this paper, the experiments based on 55 real world data sequences have been conducted to study how the translation transformation influences the grey models' error characteristics. The results show that a larger translation transformation can make the total error between the model sequence and the original sequence toward zero, and the model precision remains independent. The conclusion implies that we can use translation transformation to change the original sequence data level to simplify the model building without changing the model precision.