有协变量的推广增长曲线模型中协差阵的估计

The estimation of covariance matrix on the extensive Growth Curve Model with covariate variables

考虑协变量存在情形下推广的增长曲线模型中的参数估计问题,本文利用矩阵的谱分解和一般投影理论,在不变估计类中得到了该模型中未知协方差矩阵Σ及其线性组合tr(CΣ)的最小二乘估计.

Considering the problems of parameter estimation on the extensive Growth Curve Model when covariater Variables exist, this paper derives the least Square Estimaters of the Covariance Matrix Σ and its linear Combination tr(CΣ)if confined the class of unbias estimation. Its methods are based on the extension of projective theory and the spectral decomposition of matrix.