Publication Date

Summer 7-12-2017

Abstract

We consider standard error of the method of simulated moment (MSM) estimator for generalized linear mixed models (GLMM). Parametric bootstrap (PB) has been used to estimate the covariance matrix, in which we use the estimates to generate the simulated moments. To avoid the bias introduced by estimating the parameters and to deal with the correlated observations, (Lu, 2012) proposed a multi-stage block nonparametric bootstrap to estimate the standard errors. In this research, we compare PB and nonparametric bootstrap methods (NPB) in estimating the standard errors of MSM estimators for GLMM. Simulation results show that when the group size is large, NPB and PB perform similarly; when group size is medium, NPB performs better than PB in estimating the mean. A data application is considered to illustrate the methods discussed in this paper, using productivity of plantation roses. The data application finds that the person caring for the roses is associated with the productivity of those beds. Furthermore, we did an initial study in applying random forests to predict the productivity of the rose beds.

Degree Name

Statistics

Level of Degree

Masters

Department Name

Mathematics & Statistics

First Committee Member (Chair)

Dr. Yan Lu

Second Committee Member

Dr. Guoyi Zhang

Third Committee Member

Dr. James Degnan

Language

English

Keywords

parametric bootstrap, method of simulated moments, random forest, GLMM, generalized linear mixed models, method of moments

Document Type

Thesis

Comments

You should really read this

Share

COinS