Stat 539 - Generalized Linear Models
Spring 2017
This course will provide an introduction to the principles of generalized regression models, with an emphasis on categorical data models. Categorical data occurs extensively in both observational and experimental studies, as well as in industrial applications. The course will focus on both theory and application of methods for data analysis. Problems will be motivated from a scientific perspective. Topics covered include logistic regression, log-linear models, analysis of deviance, extrabinomial variation, quasi-likelihood, and models for correlated responses.
Upon successful completion of the course, students will be able to:
- Describe the general structure of a GLM and similarities and differences with linear models
- Estimate and interpret a logistic regression model
- Estimate and interpret a Poisson regression model
- Know of issues and some strategies for dealing with overdispersion in some GLMs
- Estimate and interpret a GLM for continuous responses that are not normally distributed
Course Information
Log into D2L course page to access course material, assignments, etc.
Data Sets
- FEV study
- Epilepsy clinical trial
- Framingham heart study
- Prostate cancer study
- Agresti horseshoe crab data
- Australia travel choices
- Alligator food choice data (comma separated; use for Homework 5 Problem 2)
- Mental impairment data
- STD data (Homework 6 Problem 3)
- Attitudes on abortion data
- Skin Cancer Prevention Study (Homework 7 Problem 4
R Code
Source into R session with code:
source("http://www.math.montana.edu/shancock/courses/stat539/r/GillenRFunctions.R")
- Jan 26: FEV case study
- Feb 2: Fitting GLMs with epilepsy data
- Feb 7: Framingham case study
- Feb 14: Berkeley graduate admissions data (R console from class)
- Feb 16:
- Nodal involvement in prostate cancer - Diagnostics for binary data
- Framingham case study (cont)
- Feb 21: Horseshoe crab data
- Feb 28: Multinomial regression with Australia travel choices data
- Mar 2: Ordinal regression with mental impairment data and housing data (R console from class)
- Mar 21: Loglinear models for contingency tables with housing data (R console from class)
- Mar 28-30: Poisson regression with gun registration law data (R console from Mar 28 and Mar 30)
- Apr 2: Zero-inflated models with horseshoe crab data
- Apr 13: GLMMs vs Marginal models with epilepsy data
- Apr 18: GLMMs vs Marginal models with attitudes on abortion data
- May 1: Bayesian logistic regression with O-ring data