Answers are grounded in this paper and cite their sources.
This paper presents a pedagogical argument to explain to quantitative geneticists how REML can be derived from maximum likelihood for estimation of variance components. The argument is first developed for N independent normal observations with mean μ and variance σ 2 and is afterward extended to a general linear mixed model structure, y ∼ N(Xβ,V). The argument is taken from expectation-maximization theory and consists of replacement of a quadratic in μ or β by its conditional expectation given the observed data and the variance components.
No takes yet. Share an insight, caveat, or question.
Jean Louis J. L. Foulley (1993) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: