Abstract:

There exists a problem that wave sequence can not be forecasted in the traditional GM(1,1) model. Based on GM(1,1) model and residual GM(1,1) model, the recurrence GM(1,1) model and the residual recurrence GM(1,1) model are established by introducing a metabolism array. After inverting adding the solution of the former model with the one of the model obtained from taking logarithm on the latter, the solution of adaptive GM(1,1) model is presented. Instance data simulation and comparison of these four methods, the results show that the adaptive GM (1,1) model has better prediction than other methods. It proposes a fundamental solution to predict wave sequence by the GM (1,1) model.