Mixed Models for Predictive Modelling in Actuarial Science

We start with a general discussion of mixed (also called multilevel) models and continue with illustrating specific (actuarial) applications of this type of models. Technical details on (linear, generalized, non–linear) mixed models follow: model assumptions, specifications, estimation techniques and methods of inference. Detailed links with useful R packages and SAS procedures are included. Illustrations in this Chapter illustrate the frequentist as well as Bayesian point of view.