A self-adaptive grey forecasting model and its application

GM(1,1) models have been widely used in various fields due to their high performance time series prediction. However, some hypotheses of the existing GM(1, 1) model family may reduce prediction cases. To solve this problem, paper proposes a self-adaptive model, termed as SAGM(1, which aims defects GM (1,1) by deleting modeling hypothesis. Moreover, novel multi-parameter simultaneous optimization scheme based on firefly algorithm is proposed, proposed adopts machine learning ideas, takes all adjustable parameters input variables, and trains it with algorithm. And Sobol' sensitivity indices are applied study global parameters, provides an important reference for parameter calibration. Finally, forecasting capability illustrated Anhui electricity consumption dataset. Results show that accuracy significantly better than other models, shown approach enhances significantly.