Nonlinear Interation Optimization Algorithm of Changing Steplength with Parameter and Its Application

摘要

In many fitting methods, after model is given,adaptability of algorithm is not ideal, although the grey model can better fit high index growth series, worse fitting big ? volatility data,and only parameter adjust to some scope.author think that the grey model is actually regarded as a linear representation of the differential,only through nonlinear representation of the differential, adaptability of model can be generalized;the other,in order to be used in discrete data, this paper combines the improved Euler interation with optimization theory ,and in an iterative process,changing steplength makes model good adaptability,getting best fitting precision.Ihe new algorithm at all overcomes shortcoming of same steplength interating.The calculating precision of examples shows this new algorithm is very effective and generalized.