线性模型回归系数的c-(K,S)型估计的小样本性质

Small-sample Properties for c-(K,S) Class Estimator in Linear Regression Model

根据线性回归模型Y=Xβ+,εE(ε)=0,COV(ε)=σ2,对回归系数的有偏估计c-(K,S)型估计进一步研究;讨论了c-(K,S)型估计的基本性质;并在均方误差阵(M SEM)准则下讨论了c-(K,S)型估计相对于最小二乘估计的优良性,有助于线性回归系数有偏估计的进一步改进.

According to the linear model Y=Xβ＋e,E（ε）=0,COV（ε）=σ^2, this paper focuses on the further studies on the c - （K,S） class estimator of the coefficients biased estimation; then gives some properties of the c - （K,S） class estimator; the superiorities of the c- （K,S）class estimator over the least square estimator is studied under the mean square error matrix （MSEM） criterion, which can improve the coefficients biased estimation.