Math 6267
Multivariate Statistical Analysis
Course Information (Spring 2007)

Time:	Tu.&Th. 12:05-1:25pm
Place:	Skiles 240
Textbook:	An Introduction to Multivariate Statistical Analysis, by Anderson, T. W. (2003), (3rd ed.) N.Y., Wiley. ISBN: 0-471-36091-0

Professor:	Serge Guillas
Office hours:	Tu.&Th. 11:00-12:00, and by appointment, Skiles 165
Phone:	(404) 894-8770
E-mail:	guillas "at" math.gatech.edu (only for important matters, please ask questions in class or during office hours)
WWW:	www.math.gatech.edu/~guillas/teaching/Sp07Math6267/Sp07Math6267index.html
	Check this website frequently for updates (homework assigments, test dates, course material..)
Homework:	Several homeworks will be assigned during the semester. You are suppose to work on them alone. They must be turned in on time, in class. Homework turned in late will not be graded.
Test & Project:	There will be one midterm (test in class, open book), and one project (to turn in at the end of the semester). You are suppose to work on the project alone. You can use your own data set and your preferred statistical/numerical software.
Grading:	Grades will be based on the homework (30%), the midterm (30%) and the project (40%).

Project deadline: Thursday May 3, at 12:30pm.

Homework:

HW1, due Feb. 13, in class

HW2, due March 15, in class.

HW3, due April 19, in class.

Midterm Exam:

Tuesday, April 3, 12:05-1:25pm, Skiles 240. Open book and lecture notes only. Bring a calculator.

Material covered in class from Chapters 1, 2, 3, 4, 5 (HW 1 & HW 2).

Project:

Maximum length: 8 pages, in 11pt or 12pt font.
Short presentation of methodology (with references) and data set.
Results should be shown using tables and figures.
Do not include code. Succintly describe the algorithm.

Project deadline: Thursday May 3, at 12:30pm, in my office (Skiles 165).

Please print the project and bring it to my office. If you can not print, please send me by e-mail a PDF document (only PDF format).

Course Goals:

Learn standard multivariate statistical analysis with rigorous mathematical derivations.
Be able to read and apply further methodology not covered in class.
Apply the techniques introduced in class to a problem that may be relevant to your graduate studies.
Carry out computations using a statistical/numerical software (e.g. R or S-Plus, SAS, Matlab) in this area.

Tentative Syllabus:

Introduction.
The Multivariate Normal Distribution. [Review]
Estimation of the Mean Vector and the Covariance Matrix.
The Distributions and Uses of Sample Correlation Coefficients.
The Generalized T²-Statistic. Application to quality control.
Classification of Observations.
The Distribution of the Sample Covariance Matrix. The Wishart Distribution.
Testing the General Linear Hypothesis. Multivariate Analysis of Variance.
Principal Components. (Application to Geophysical Sciences: EOFs)
Canonical Correlations and Canonical Variables. Application to downscaling.
Factor Analysis

Extensions to the Infinite-Dimensional Case: Functional Data Analysis [if time permits]