Decomposition of multivariate statistical models

Eva Fišerová, Lubomír Kubáček

Abstract


A decomposition of a multivariate model into a system of two
simpler multivariate models is considered. It is proposed a test for decision which
of models should be used in order to obtain more efficient estimators. For a case,
when a part of parameters is known, it is given a tolerance domain such that if the
parameters lie within this domain, then it is better to neglect these parameters.

Full Text:

PDF


DOI: https://doi.org/10.2478/tatra.v51i1.149