Outliers on leave-one-out cross-validated PLS1 regression models. Extending PLS1 to PRLS.

This article is mainly concerned with the effects of outliers on model selection and assessment of PLS1 regression models, where PLS1 corresponds to PLS regression on one response. The leave-one-out cross-validation has been used in order to select the final regression model and assess its performance. PLS1 proves to be sensitive to out-liers. Therefore, a modified and robust algorithm, the Partial Reweighted Least Squares (denoted as PRLS), has been applied. PRLS algorithm reduces the dimension of the regression problem and improves model’s performance