The course covers the
basic theory of the multivariate normal distribution
and its application to multivariate inference about a single mean,
comparison of several means and multivariate linear regression. As time and
interest permit, further related topics may also be covered.
We will study methods of analysis for data which consist of
observations on a number of variables. The primary aim will be
interpretation of the data, starting with the multivariate normal
distribution and proceeding to the standing multivariate inference
theory based on linear models. Sufficient theory will be developed to
facilitate an understanding of the main ideas. This will necessitate a
good background in matrix
algebra, and some knowledge of vector spaces as well. Computers will be
used extensively, and familiarity with elementary use of SAS or S+
will be assumed. Topics covered will include the multivariate
normal population, inference about means and covariance, multivariate
linear models, principal component analysis, , and some discussion of
canonical correlation analysis, discriminant and classification, factor
analysis and cluster analysis, as time permits.
Grades will be based on a combination of class quizzes and a final
examination, plus homework including a group project. The coordinator
may permit students to enrol who have background "equivalent to" the
formal prerequisites below.
Prerequisite:AS/SC/AK/MATH 3131 3.0;
AS/SC/MATH 3034 3.0 or AS/SC/MATH 3230 3.0;
AS/SC/MATH 2022 3.0 or AS/SC/AK/MATH 2222 3.0.
Coordinator: M. Asgharian
