On Predictors and Estimators under a Constrained Partitioned Linear Model and its Reduced Models

Authors : Melek ERİŞ BÜYÜKKAYA, Nesrin GÜLER

Pages : 192-200

Doi:10.33401/fujma.1134600

View : 29 | Download : 8

Publication Date : 2022-09-23

Article Type : Research Paper

Abstract :In this study, we consider a partitioned linear model with linear partial parameter constrains, known as a constrained partitioned linear model insert ignore into journalissuearticles values(CPLM);, and its reduced models. A group of formulas on best linear unbiased predictors insert ignore into journalissuearticles values(BLUPs); and best linear unbiased estimators insert ignore into journalissuearticles values(BLUEs); in CPLM is derived via some quadratic matrix optimization methods, and further many basic properties of the predictors and estimators are established under some general assumptions. Our main purpose is to derive various inequalities and equalities for the comparison of covariance matrices of BLUPs and BLUEs under CPLM and its reduced models.
Keywords : BLUP, Correctly reduced models, Covariance matrix, Rank