Canonical Correlations and Generalized SVD (Singular Value Decomposition): Applications and New Algorithms

摘要

This paper considers canonical correlations and a generalization of the singular value decomposition (SVD) that involves three matrices. The authors show how the two matrix problems are related and how they can be used in important applications such as weighted least squares and optimal prediction. They present two new computational procedures for the problems based on implicit SVD methods for triple matrix products. These algorithms are well suited for parallel implementation. Keyword: Reprints. (KR)