The least square estimate of covariance matrices in the growth curve model with random effects
Gui Yong-xin · Journal of Central China Normal University · 2004
In this paper, a least square estimate of the covariance matrices Σ, Γ and their linear function tr(CΣ+DΓ) of the growth curve model with random regression (coefficient) structure has been given respectively by using the projective theory and the spectrum decomposition of matrix. Finally, it gives some optimal properties of tr(CΣ\+*+DΓ\+*).