Estimation and accuracy evaluation of variance and covariance components based on more general functional model①
(Institute of Surveying Engineering and Land Information, Central South University of Technology,Changsha 410083, P.R.China)
Abstract: Started from the more general functional model and based on the work of Koch K R (1986) and QU Zi-qiang (1989), marginal likelihood function of variance and covariance components is derived and is identical with the orthogonal complement likelihood function. Minimum norm quadratic unibiased estimator (MINQUE) is developed, which expands the formula by Rao C R (1973). It is proved that Helmert type estimation, MINQUE, BQUE(Best quadratic unibiased estimation) and maximum likelihood estimation are identical with one another. Besides, a universal formula for accuracy evalution is presented. Through these work, a universal theory of variance and covariance components is established.
Key words: variance component; marginal likelihood; accuracy evaluation