用Simpson公式构造背景值的GM(1,1)建模新方法

提出了用Simpson数值积分公式构造背景值的GM(1,1)建模新方法,并通过算法分析和一些实例说明了该方法对很多时间序列应用方便,且其模型的拟合精度也比一些文献中的建模方法有明显改进.认为所提出的方法是建立GM(1,1)预测模型时值得考虑的一个新方法,这不仅将对GM(1,1)建模方法的理论研究提供必要的算法依据,而且...

Constructing the Background Value for GM(1,1) Model Based on Simpson Formula

A new method was proposed to construct the background value for GM（1,1） model by using the Simpson numerical integration formula. The algorithm analysis of some examples indicates that this new method is adaptable to many time series, and the fitting precision is improved considerably compared with that in other literatures. So we think our method de- serves to be considered when one sets up GM（1,1） prediction model. The method not only provides necessary algorithm basis for the theory of GM（1,1） model, but also has some reference value for properly using GM（1,1） prediction model.