ESTIMATION AND INFERENCES IN LINEAR MIXED EFFECTS MODELS: A COMPARATIVE STUDY

Smoothing methods that use basis functions with penalization can be formulated as fits in formlinear mixed effects models. This allows s uch models to be fitted using sta ndard mixed models structures. In this paper we provide an estimation and inference for linear mixed models using restrict- ed maximum likelihood and penalized spline smoothing, and describe the connection between the two. To this end, a real data example is considered and model is fitted in R using diff erent package. We see that penalized spline smoothing expressed in form of linear mixed model gives the better results than standard mixed effects model.