Class Information
Lecture Notes
- Introduction (1)
- Exploratory Data Analysis (2 hrs)
- General Linear Models (3 hrs)
- General Linear Models: Case Study (2 hrs)
- Linear Mixed Models (2 hrs)
- Linear Mixed Models: Case Study (2 hrs)
- Generalized Linear Models, Quasi-likelihood and Estimating Functions (4
hrs)
- GEE Variants and Case Studies (3 hrs)
- Likelihood Models for Repeated Binary Data (2 hrs)
- Modeling Approaches: Marginal, Random Effects and Transition Models (2
hrs)
- Generalized Linear Mixed Models (5 hrs)
- Transition Models and Marginalized Models (1 hr)
- Time-Dependent Variables (2 hrs)
- Missing Data in Longitudinal Studies (2 hrs)
Data Sets
Homework Assignments
Projects
Links
Software
R and S-PLUS
- Linear Mixed Models (appendix to John Fox's An R and S-PLUS Companion to
Applied Regression).
- R packages available at CRAN:
nlme, geepack, gee, lme4, Design, lgtdl,
longitudinal.
- V. Carey's packages:
alr (Alternating Logistic Regression) and yags (Yet
Another GEE Solver),
- Jim Lindsey's packages:
repeated (non-normal repeated measurements models),
gnlm (generalized non-linear models).
- NLME (S-PLUS)
- Ord.EE: GEE for correlated ordinal data (S-PLUS). The ordgee function in
package
geepack does the same thing.
SAS
Others
I haven't tried any of these.
- MPH.fit: ML fitting of multinomial-poisson homogeneous (mph)
models for contingency tables (JB Lang, U of Iowa)
- VGAM: vector generalized linear and additive models.
- Negative Binomial Loglinear Mixed Model: (Microsoft Windows only)
- MAREG and WinMAREG: marginal regression models for correlated using GEE or
ML.
- EE: a Windows program that implements Prentice and Zhao's GEE model.