UGA Bulletin Logo

Linear and Generalized Linear Models


Course Description

Extension of linear models to data with heterogeneous variance and correlated repeated measures. Mixed effects, log-linear, and generalized linear models are considered. Statistical inference is carried out using quasi-likelihood, conditional likelihood, marginal likelihood, and generalized estimating equations. Applications in public health and biomedical research.


Athena Title

LINEAR AND GLM


Prerequisite

BIOS 8010


Semester Course Offered

Offered spring


Grading System

A - F (Traditional)


Course Objectives

A student completing this course should be able to: 1. Describe mixed effects, log-linear and generalized linear models. 2. Describe various approaches to model fitting including maximum likelihood, conditional likelihood, marginal likelihood, quasi-likelihood, and generalized estimating equations. 3. Describe the inferential properties of parameter estimators. 4. Fit linear and generalized linear models using statistical software. 5. Apply linear and generalized linear models to public health and biomedical data.


Topical Outline

1. Background for generalized linear models a. Linear models and likelihood theory b. Exponential family of distributions c. Motivation: Why more than linear regression 2. Introduction to generalized linear models a. Formulation / Link functions b. Maximum likelihood estimation and inference c. Deviance 3. Generalized linear models for correlated continuous data a. Models and estimation b. Linear mixed models c. Longitudinal data analysis 4. Analysis of discrete and counts data a. Logistic regression, probit regression, complimentary log-log models b. Poisson regression, negative binomial regression c. Log-linear models 5. Correlated discrete data a. Marginal and conditional models b. Generalized linear mixed models (GLMMs) c. Conditional likelihoods for GLMMs d. Overdispersion 6. Model diagnostics a. Diagnostic measures b. Local influence analysis c. Goodness-of-fit statistics 7. Generalized estimating equations a. Estimating functions b. Quasi-likelihood c. Generalized estimating equations


Syllabus


Public CV