函数变换提高灰色预测模型精度的条件  

Conditions of Function Transformation Improve the Accuracy of Grey Prediction Model

作　　者：李玻[1] 方玲[1] 

机构地区：[1]后勤工程学院基础部

出　　处：《后勤工程学院学报》2009年第4期86-90,共5页Journal of Logistical Engineering University

基　　金：国家自然科学基金资助项目(60576048);后勤工程学院学术创新基金资助(YZ07-41101)

摘　　要：指出选择函数变换来提高模型精度主要与三个方面的因素有关：提高数据序列的光滑比、调整数据序列的级比和确保凸凹性。在调整级比方面，扩充了级比压缩变换的相关定理，并讨论了几种常用的函数变换在上述三个方面的关系。最后结合实例提出，对于变换后的数据序列若不满足这三方面的要求可以再进行函数变换以达到提高模型精度的目的。This text indicates that using selecting function transformation to improve the accuracy of model relates to three factors： increasing the smooth ratio of original series, adjusting the stepwise and maintaining the convex and concave character. In the aspect of adjustment stepwise,the theorem of stepwise compression transformation is expanded and the relationship of several commonly used function transformations in these three aspects is discussed. Finally, it puts forward that if date series after transformation does not meet these requirements it can further use function transformation with a view to increase the accuracy of the model.

关 键 词：灰色预测模型 光滑比 级比压缩变换 凸凹性 

分 类 号：O21[理学—概率论与数理统计]