On Translational Operator-based DGM(1,1) Model and its Properties

摘要

Function transformation is a common approach to improve accuracy of the grey forecasting model. Discrete GM(1,1) model (DGM(1,1)) is an important member of the family of grey system models. To improve the fitting accuracy of an oscillating sequence, the variable speed translational operator-based DGM(1,1) model was built and the fitting accuracy was compared with those of other methods in this study. Furthermore, to determine how variable speed transitional operator affect predictive accuracy of the DGM(1,1) model, the empirical analysis of property for variable speed translational operator was conducted. The results demonstrated that the proposed variable speed transitional operator to reduce second-order stepwise variance and second-order stepwise ratio deviation were effective in improving accuracy of DGM(1,1) model for the oscillating sequence. It will supplement the criterion of function transformation and widen the application range of the grey forecasting model.