Fast partial quantile regression

Partial least squares (PLS) is a dimensionality reduction technique used as an alternative to ordinary (OLS) in situations where the data colinear or high dimensional. Both PLS and OLS provide mean based estimates, which are extremely sensitive presence of outliers heavy tailed distributions. In contrast, quantile regression that computes robust estimates. this work, multivariate extended framework, obtaining theoretical formulation problem we call fast partial (fPQR), provides An efficient implementation fPQR also derived, its performance studied through simulation experiments chemometrics well known biscuit dough dataset, real dimensional example.