回归系数的一类线性估计相对于GLS估计的优良性

摘要

对线性回归模型中的一类线性估计,在均方误差矩阵准则和PC准则下,研究了它相对于广义最小二乘估计的优良性.当设计阵为非列满秩时,讨论了回归系数的可估函数的优良性.%The superiorities about a class of linear estimator of regression coefficients over generalized least square (GLS) estimator were studied under the mean square error matrix (MSEM) criterion and Pitman closeness (PC) criterion. The superiorities of the linear estimator of the estimable function for regression coefficients were shown in a non-fullrank case.