摘要
在原有的等距化GM(1,1)模型的基础上,重新给出一个建模机理较简单、形式规范、使用方便的等距化GM(1,1)模型和选用这种模型的条件.这种模型与回归模型相比有较多的优点,主要表现为在一定的条件下,只要取这一个规范样式来建模即可.而在同一条件下可供选择的回归模型的样式往往有多个,取那一个为最优,需经逐个建模后的检验比较.即使是选出的最优回归模型,其拟合与预测精度时常也不及规范样式的等距化模型.本文用实例说明了这一点.
On the basis of a ready equal-distantified model GM (1, 1), this article gives a new model GM(1, 1) with a simpler modeling mechanism, standard form and a convenient usage, and the choosing conditions of the model to be used. Compared with the routine model, it has more advantages: mainly, under the given conditions, it is enough only to take this standard form for modeling. while under the same conditions the forms of the routine model can often be selected. The best of them is determined on the test, and all of the modeling forms are compared one by one. It is clear that the former is simple and the latter troublesome. Even if the best routine model has been chosen, its combination and prediction accuracy are often not so good as the standard form of the equal-distantified model. The examples can illustrate this.
出处
《石油大学学报（自然科学版）》
CSCD
1993年第1期114-120,共7页
Journal of the University of Petroleum,China(Edition of Natural Science)
