Linear Mixed Model in the Light of Future Data
Kunio Takezawa *
Agroinformatics Division, Agricultural Research Center, National Agriculture and Food Research Organization Kannondai 3-1-1, Tsukuba, Ibaraki 305-8666, Japan.
*Author to whom correspondence should be addressed.
Abstract
The maximum likelihood and restricted (or residual) likelihood methods are common tools for estimating variances in linear mixed models. However, regression in the light of future data can yield different results. Investigations into the characteristics of this new variance are expected to promote the effective use of data in fields such as ecology and genetic statistics. Our numerical simulations show that the estimates of variances in the light of future data are substantially different from those given by the maximum likelihood and restricted (or residual) likelihood methods.
Keywords: Expected log-likelihood, linear mixed model, maximum likelihood estimator, optimization, third variance.