广义G-M模型参数的Bayes线性无偏估计及其优良性

On Superiority of Bayes Linear Unbiased Variance Estimator in Generalized Gauss-Markov Model

针对广义Gauss-Markov(G-M)模型,采用Bayes估计方法获得参数的Bayes线性无偏估计(BLUE),在均方误差矩阵准则下与广义最小二乘(GLS)估计进行比较,导出了4种相对效率的界,讨论了在PC准则下BLUE相对于GLS估计的优良性.

The estimation method and properties of unknown parameters have been studied in the general- ized Gauss-Markov（G-M）model. The Bayes linear unbiased estimators （BLUE） of parameters have been derived in Bayes estimation method. Then we have compared the Bayes estimator with generalized least square （GLS） estimator in parametric part under the mean square error matrix criterion conditions. Fur thermore, we have obtained the bounds of four relative efficiencies, respectively. In addition, the superior ity of the BLUE has been studied with respect to GLS estimator in terms of the PC criterion.