混合模型的一种新的相对效率

A new relative efficiency in mixed model

研究了带附加信息的混合模型回归系数的估计,在考虑混合估计和LS估计协方差阵之间进一步联系的基础上,给出度量混合估计和LS估计差异的相对效率的一种新定义,并利用矩阵不等式分别求出新相对效率的上、下界.

In this paper, mixed estimator of regression coefficients under prior information is considered. On the basis of considering the further relation between the covariance matrix of mixed estimator and the covariance matrix of LSE, a new relative efficiency of mixed estimator with respect to LSE is given. The upper and lower bounds of the new relative efficiency are obtained by using matrix inequality.