Nonparametric mixed -effect models

Ping Ma, Purdue University

Abstract

Mixed-effect models are widely used for the analysis of correlated data such as longitudinal data and repeated measures. In this thesis, we study an approach to the nonparametric estimation of mixed-effect models and generalized mixed-effect models. We consider models with parametric random effects and flexible fixed effects, and employ the penalized likehood method to estimate the models. The issues to be addressed are efficient computation methods and the selection of smoothing parameters through cross-validation methods, which is shown to yield optimal smoothing for both real and latent random effects for Gaussian responses. Simulation studies are conducted to investigate the empirical performance of various cross-validation techniques in the context. Real data examples are presented to demonstrate the applications of the methodology.

Degree

Ph.D.

Advisors

Gu, Purdue University.

Subject Area

Statistics

COinS