Alternate Procedures for Analysis of Multivariate Regression Models

Abstract

In previous chapters, we have developed various reduced-rank multivariate regression models, and indicated their usefulness in different applications as dimension-reduction tools. We now briefly survey and discuss some other related multivariate regression modeling methodologies that have similar parameter reduction objectives as reduced-rank regression, such as multivariate ridge regression, partial least squares, joint continuum regression, and other shrinkage and regularization techniques. Some of these procedures are designed particularly for situations where there is a very large number n of predictor variables relative to the sample size T including, for example, n > T.