基于Cp统计量的自变量选择原则

The Independent Variable Selection Rule Based on the Cp Statistics

对回归模型ｙ＝β０＋β１ｘ１＋…＋βｔｘｔ（其中ｙ是ｑ×１随机向量，βｉ为ｑ×１维参数向量），提出在ｑ≥１的情况下，基于Ｃｐ统计量的自变量的选择原则：选择自变量是子集Ｐ，使其相对应的Ｃｐ值满足条件ＣＰ≤γα（ｔ，ｎ－ｔ－１，ｑ）．

For the regression model y=βo+ β1x1+…+β1x1, where y is q ×1 random vector and βiis q ×1 parameter vecter, the independent variable selection rule based on the CP statistics under the condition of β≥ 1: The selection of the independent yariable subset P can be done through the calculation of Cp corresponding to P, which satisfied CP≤γα(t,n-t-1,q).