提出了反余弦函数变换方法,证明了这种变换是级比压缩变换,能够提高序列光滑度,可以保持序列凹凸性,不会增大还原误差,满足数据变换的构造准则.通过具体算例表明,基于反余弦函数变换的GM(1,1)模型的预测精度优于传统GM(1,1)模型和基于幂函数变换的GM(1,1)模型,说明了该变换的有效性.
Arccosine function transformation is proposed and proved to be stepwise compression transformation,and this kind of transformation can also remain the concave-convex of original sequence and decrease the reduction error.That is,it meets the guidelines for the construction of the data transformation.An example is given to demonstrate the optimized GM(1,1) based on arccosine function transformation is better than the traditional GM(1,1)and the improved GM(1,1) based on power function transformation,and thus this data transformation is efficient.