Statistics | Multivariate Methods II
S481 | 15320 | Steen Andersson

This course requires knowledge in basic statistical theory, in
particular the basics in multivariate statistical analysis. The actual
content of this course depends on the audience and the teaching method
depends on the number of students in the class. The course material
will be published on the web. Thus, no textbook is required. The
topics will, as a rule, be taught in depth and detail. The content
will be chosen from the following list of topics/subtopics:

(I) Advanced multivariate models: (i) Conditional models, (ii)
Covariance analysis, (iii) Growth curve models, (iv) Symmetry normal
models, (v) Graphical normal/discrete models, (iv) Missing data normal
models, (v) Mixture models.

(II) Eigenvalues: (i) Canonical correlations, (ii) Principal component
analysis, (iii) Factor analysis models, (iv)
Discriminant/classification analysis, (v) Testing and eigenvalues in
multivariate normal models.

(III) Exponential families: (i) Generalized linear models, (ii)
Multivariate logistic regression, (iii) Rasch measurement models, (iv)
Conjugated priors, (v) EM/scorings algorithm, (vi) Testing in
exponential families, (vii) Models in the Wishart distributions.

References: A.J. Izenman (2008) Modern Multivariate Statistical
Techniques, Springer.

Brian S. Everitt (2005) An R and S-Plus Companion to Multivariate
Analysis, Springer.