A weighted view on the partial least-squares algorithm (Q1977038): Difference between revisions

This paper provides deep insight into the Partial Least Squares (PLS) Algorithm and gives a new, non-iterative formulation of the PLS algorithm. The relation between the PLS, the Krilov controllability matrix and the weighted least squares are studied in detail. Both the univariate and the multivariate cases are considered. A new PLS algorithm for multivariate data is proposed.