EXST 7037 - Multivariate Statistical Data Analysis

Fall 2005

Course Description: EXST 7037, Comparison of multivariate techniques and analyses; emphasis on discriminant analysis, factor analysis and principal component analysis, canonical correlation, cluster analysis, and multivariate analysis of variance and repeated measures.

Course Prerequisites: Statistical Methods II (EXST 7013, 7014, 7015) or equivalent (must check with Instructor); and knowledge of matrix algebra at a basic level. The prerequisite material is essential as this course depends upon the estimation, hypothesis testing, regression, analysis of variance, model diagnostics, and experimental design concepts covered in the statistical methods sequences (7003-7013, 7004-7014, and 7005-7015). If you have not had the prerequisite courses, you should take those first.

Course Objectives: General objectives include the following:

  • Help student to recognize multivariate data and situations when multivariate analysis is necessary or may prove beneficial.
  • Familiarize student with the terminology and mathematical language used to describe multivariate techniques.
  • Introduce student to several of the more "well known" multivariate methods.
  • Demonstrate software that can be used to perform most of the multivariate analyses discussed.
  • Provide hands-on practice with multivariate techniques that are often employed in the preparation of theses, dissertations, and research papers.
  • Provide a stimulus for the student to take a greater role in her data analysis and stimulate the student's general interest in statistics.

Instructor: Dr. E. Barry Moser, Professor and Head, Dept. Experimental Statistics

Office:

149A Ag. Administration Bldg.

Phone:

225-578-8376

Email:

bmoser@lsu.edu --- Use BLACKBOARD for questions about course material.

Web:

http://www.stat.lsu.edu/faculty/moser/exst7037/

Office Hours: Tu 1:30p-2:30p, Th 10:35a-11:30a, or by appointment

Text: Johnson, R. A. and D. W. Wichern. 2002. Applied Multivariate Statistical Analysis, Fifth Edition. Prentice Hall, Upper Saddle River, NJ, 767pp.; Course Examples and Selected Notes, available from Copies Too on Nicholson near Lee Drive.

Assignments and Grading: Grades will be based upon 3 monthly exams, 1 group poster project, and a final examination (equal weight to each of the 5 items). Exams must be taken at the assigned times unless prior approval from Dr. Moser is obtained. Note, taking the final exam early is not permitted. Unauthorized/unexcused missed exams will count as 0 points. You are to do your own work on all exams. Project work is to be solely a product of the group team, and each team member is expected to contribute significantly to their project. Issues of academic misconduct will be referred to the Dean of Students.

Grading Scale and Event Dates: Letter grade assignments will be made according to the scale below. Exam and poster due dates are also given. An exam or due date may be moved depending upon our progress in the course, or to accommodate my other work responsibilities. Any changes in exam or due dates will be announced in class and/or via Blackboard. Check your Blackboard account regularly.
 
 

Score

Minimum
Grade

                                    

Event

Date

90-100

A

 

Exam I

Sep 29

80-89.9

B

 

Exam II

Oct 27

70-79.9

C

 

Exam III

Nov 22

60-69.9

D

 

Poster

Dec 6

0-59.9

F

 

Final
Exam

Dec 17
12:30p-2:30p

CONTENT OUTLINE

 
 

Class
Period1

Topic

Readings2

1

Introduction, Objectives, Multivariate Data, Matrix Algebra and Vector Spaces

1-30,
50-111

2

Statistical Distance, Expected Values, Variances and Covariances of Linear Combinations, Sample Geometry

30-37
67-79, 112-148

3-4

Multivariate Normal Distribution

149-209

5-6

Principal Components Analysis, Biplots

426-458, 719-723

7-8

Cluster Analysis (Hierarchical and Non-Hierarchical)

668-700

9

Exam I

 

10

Multidimensional Scaling, Principal Coordinates Analysis

700-708

11-12

Inferences about a Mean Vector, Hotelling's T2

210-219

13

Confidence Regions and Simultaneous Comparisons, Missing Data

220-238, 252-256

14

Two-sample T2

272-293

15

Introduction to MANOVA

293-305, 395

16

Exam II

 

17-18

MANOVA and Linear Models, Compositional Data Analysis3

305-323, 327-332, 
354-410.

19-20

Profile and Repeated Measures Analysis

272-282, 318-327

21-23

Discrimination and Classification, Canonical Discriminant Analysis, Data Mining

581-628, 641-646,  
628-641, 731-747

24

Exam III

 

25-26

Factor Analysis and the Factor Model, Factor Rotation, Scores, Strategy

477-524

27

Path and Structural Equation Models

524-529

28

Correspondence Analysis3, Procrustes Analysis3, Canonical Correlation Analysis3

709-719, 723-730,
543-580

1Class periods are approximate. Exam dates are subject to change.
2Johnson, R. A. and D. W. Wichern. 2002. Applied Multivariate Statistical Analysis, Fifth Edition. Prentice Hall, Upper Saddle River, NJ, 767pp.
3Time permitting, otherwise covered within other topics.

To Top of Document