Wenxue Ge

The Prediction of Random Effects in Generalized Linear Mixed Model

The introduction of random effects in Generalized Liner Mixed Model, or GLMM, raises the topic about the prediction of random effects. Best linear unbiased prediction, or BLUP, is one of the methods of predicting random effects in linear mixed model. This method could generate the unbiased estimations with the minimum mean squared errors. BLUP can be derived by maximizing the joint density function under the normality assumption. The procedures are similar to the maximum likelihood method processes. The concept of Bayesian and posterior density also contributes to deriving the BLUP when the distributions are specified.