Estimation of variance components in random models
On estimation of parameters in linear models
Quadratically admissible estimators in random models.
The author is concerned with a random model of the form y=1μ+X1β1+⋯+Xkβk, where 1 is the n-vector of ones, μ is an unknown constant, Xi is an n×pi matrix of known constants, and βi is a pi-vector of random variables. It is assumed that pk and Xk=I the identity matrix of order n, and also that β1 are mutually independent, with βk∼Npi, i=1,⋯,k, where Vk=I, Vi is a known pi nonnegative definite matrix, and σi is an unknown nonnegative parameter, with the additional requirement that σk be strictly...
O I Konferencji ze Statystyki Matematycznej w Wiśle
.
Best linear plus quadratic unbiased estimation of parameters in mixed linear models
Page 1