STSCI 7190 - Advanced Multivariate Statistics

Spring. 4 credits. Student option grading (no audit).

Prerequisite/Corequisite: ORIE 6700 or MATH 6730 (or equivalent) and at least one course in probability, or permission of instructor.

S. Basu.

Structure learning and prediction of large, complex systems arising in modern biological and social sciences require multivariate statistical methods that are computationally efficient, perform dimension reduction and are able to capture nonlinear relationships. This course will focus on optimization and statistical aspects of selected multivariate data analysis techniques. Topics in dimension reduction will include principal component analysis, ridge regression, as well as modern sparsity-based methods for graphical modeling and large-scale systems identification from time series data. Topics in nonlinear modeling will include kernel methods, decision trees and their ensembles. Applications in genomics, neuroscience and financial economics will be considered. Depending on time and interest, some additional topics in clustering may also be covered. The R programming language will be used for implementing statistical methods.