A Generalized Linear Mixed Model with Normal Mixture Random Effects

• Activity Number: 696
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Lanfeng Pan* and Yehua Li and Kevin He and Yi Li
• Companies: Iowa State University and Iowa State University and University of Michigan and University of Michigan
• Keywords: Generalized Linear Mixed Model ; Monte Carlo EM ; Normal Mixture ; Number of Components ; Consistency

Abstract

We consider generalized linear mixed model with random effects being modeled as normal mixture. We use normal mixture to address the problem of non homogeneous random effects. Although the assumed distribution of random effects is not important for the estimation of covariates coefficients, it does have an influence on the prediction of the random effects. And we propose a likelihood ratio test to select the number of components for the mixture. The method is applied to evaluate the performance of organ transplant centers in United States. Simulations show GLMM with normal mixture can capture the non-homogeneity.