Classification and Estimation of Growth Curves, a Bayesian Approach,

摘要

Classification of observations and estimation of parameters in growth curves are considered from a Bayesian viewpoint. The author considers two different covariance matrices, (1) Sigma arbitrary positive definite and (2) Sigma = X(Gamma)(X prime) + Z theta(Z prime), which was first considered by Rao, where X is p x m of rank m < p and Z is p x p - m such that (X prime)Z = 0.