The Study of Relative Efficiency in Growth Curve Model

Among them, Y is the n q  random observation matrix; e is the n q  random error matrix, , nm kq X Z is the design matrix that we are known, and   r X m  ,   r Z k  , B is the p k  regression parameter matrix that we are unknown, V and  are positive definite matrices with the size q q  and n n  -order respectively.  expresses product of matrix Kronecker. e expresses a column vector of the e straightened by column. In the model of (1), there are two more common kinds in the B estimator class [1], [2]: One is the best linear unbiased estimator (BLUE), recorded as: