CAS MA 576: Generalized Linear Models (Spring 2009)


Instructor: Dr. Surajit Ray
Department of Mathematics and Statistics
MCS 222, 111 Cummington Street, Boston, MA 02215
Phone: (617) 353-5209, Fax: (617) 353-8100
Class Time: Tue,Thu 9:30am 11:00am
Class Room: MCS 148
Office Hours: TBD

Blackboard Login ( BU Students)



Course Description:
Covers topics in linear models beyond MA 575: generalized linear models, analysis of binary and polytomous data, log-linear models, multivariate response models, non-linear models, graphical models and relevant model selection techniques. Additional topics in modern regression as time allows.



Grading:
Homework 30%, Midterms 30%, Final 30%, Class participication and discussions 10%.
Biweekly homeworks and 2 Midterms.



Week-by-Week Syllabus:


Week Topic
1 Summary of Topics covered in MA 575: Linear models.
2 Outline of Generalized Linear Models (GLM).
3 Non-normal distribution and Link functions.
4 Models for Continuous Data with Constant Variance.
5 Estimation in GLM: MLE, Extension of Least Square estimation.
6 Models for Binary Data Estimation and Deviance measures.
7 Models for Polytomous Data.
8 Log-linear Models.
9 Conditional likelihood and Quasi Likelihood functions.
10 Conditional likelihood and Quasi Likelihood functions.
11 Models with non-linear parameters.
12 Model Checking and Model Selection for GLM's.
13 Modeling the Variance: Structural Covariance, Structural Equation Model and Multivariate response models.
14 Introduction to Robust Regression and Principal Component Regression.



Textbook:
P. McCullagh; John A. Nelder Generalized Linear Models, Second Edition Chapman and Hall

List Price: $104.95
Cat. #: C1760
ISBN: 0412317605
Publication Date: 8/1/1989
Number of Pages: 532
Availability: In Stock

Supplementary Book
Charles E. McCulloch, Shayle R. Searle Generalized, Linear, and Mixed Models Wiley, New York

Generalized, Linear, and Mixed Models [E-Book]
Charles E. McCulloch, Shayle R. Searle
ISBN: 0-471-65404-3
E-Book
April 2004
US $120.00 Purchase This E-Book
Software
We will use the statistical software R (or its commercial predecessor, S-Plus) for computing in this course. The software is now becoming a standard for users both in
and outside of statistics, particularly those that need to do programming of new methods. The software is free and compatible with Windows, Mac, and Linux/Unix. It may be downloaded from cran.us.r-project.org. The main website for the R project is www.r-project.org. For those that do not wish to install R on their own machines, it can be accessed in a command-line version on the ACS machines and, for those students in the department, on the /machines in the Department of Mathematics and Statistics.

It is expected that all students will be responsible for getting up to speed on the
basics of the R package on their own within the second week or so.  There are some excellent resources on the web e.g the official manuals and the R Primer . During the course of the semester I will list more resources.

Please Note:
You are responsible for knowing, and abiding by, the provisions of the GRS Academic Conduct Code, which is posted at
http://www.bu.edu/grs/academics/resources/adp.html
Violations of the code are punishable by sanctions including expulsion from the University.