通过线性充分和线性完全统计量比较带约束的一般Gauss-Markov模型

Comparison of Restricted General Gauss-Markov Model by Linearly Sufficient Statistics and Linearly Complete Statistics

对于带有约束的一般Gauss-Markov模型L_1(?){Y_1,X_1β,V_1|L_1β=P_1},L_2(?){Y_2,X_2β,V_2|L_2β=P_2}通过线性充分统计量、线性完全统计量、最小线性充分统计量,我们刻划了L_1和L_2的如下的序的关系:L_1优于L_2,L_1不劣于L_2,L_1等价于L_2。

Abstrct For the two restricted general Gauss-Markov model LZ = (F1, X1B, V1|L1B=p1>, and L2={Yz, X2B,V2|L2B= p2}. By linearly sufficient statistics, linearly minimal sufficient statistics and linearly complete statistics, we characterized the following relations. L1 at least as good as L2, L1better than L2,L1 equivalent to L2