一个推广增长曲线模型协差阵的最小二乘估计及其优良性

The Least Square Estimate of Covariance Matrix in an Extended Growth Curve Model and its Optimal Property

在准正态情形与独立同分布情形下,分别给出了一个推广增长曲线模型中协差阵 的最小二乘估计,并得到了最小二乘估计成为一致最小方差不变二次无偏估计的充要条件.

Let Y = be an extended growth curve model, where ε= (ε(1), …,ε(s))' is a random matrix. This paper gives the LSE of tr(CΣ) in the case of quasi-normal distribution or i.i.d, and meanwhile, the necessary and sufficient condition is obtained for the LSE to be UMVIQUE.