| Course ID: | STAT 4250/6250. 3 hours. |
| Course Title: | Applied Multivariate Analysis and Statistical Learning |
Course
Description: | The methodology of multivariate statistics and machine learning
for students specializing in statistics. Topics include inference
on multivariate means, multivariate analysis of variance,
principal component analysis, linear discriminant analysis,
factor analysis, linear discrimination, classification trees,
multi-dimensional scaling, canonical correlation analysis,
clustering, support vector machines, and ensemble methods. |
| Oasis Title: | Applied Multivariate Analysis |
| Undergraduate Prerequisite: | (MATH 3300 or MATH 3300E or MATH 3000) and (MATH 2270 or MATH 2270H or MATH 2500 or MATH 2500E) and STAT 4230/6230 and (STAT 4360/6360 or STAT 4360E/6360E or STAT 4365/6365) |
| Graduate Prerequisite: | STAT 6420 or permission of department |
Semester Course
Offered: | Offered every year. |
| Grading System: | A-F (Traditional) |
|
| Course Objectives: | Students will learn how to visualize multivariate data. Students
will learn some basic matrix algebra for statistical use and be
able to use it to understand the methods. Students will learn
multivariate analogs of one and two population inference on
means, and the multivariate analysis of variance. Students will
learn how to determine which multivariate methods are
appropriate for a given situation. Students will learn the basic
logic behind each method. For each method, students will learn
the underlying assumptions and conditions under which the
methodology can be expected to perform well, the type of
questions it addresses, the results it yields, and the proper
interpretation of those results. Students will learn how to
implement the methods covered in the course using appropriate
statistical software. |
| Topical Outline: | Relevant vector and matrix algebra, basic summary statistics for
multivariate data, the multivariate normal distribution,
inference for multivariate means, multivariate analysis of
variance (MANOVA), principal components, exploratory factor
analysis, discriminant analysis, canonical correlation analysis,
tree-based methods, support vector machines, and cross-validation. |