Sequential and orthogonalized Partial Least Squares (SO-PLS)

SO-PLS is a multi-block regression method (which can also be extended to classification problems through the combination with linear discriminant analysis), having the characteristic that PLS models between each block of predictors and the response(s) are sequentially calculated, after orthogonalization with respect to the scores of the previous regressions.

More details can be found in the following references:

A. Biancolillo, T. Næs, The Sequential and Orthogonalized PLS Regression for Multiblock Regression: Theory, Examples, and Extensions. In: M. Cocchi (Ed.), Data Fusion Methodology and Application, Data Handling in Science and Technology vol.31, Elsevier, Oxford, 2019, 157-177.

A. Biancolillo, I. Måge, T. Næs, Combining SO-PLS and linear discriminant analysis for multi-block classification, Chemometr. Intell. Lab. Syst. 141 (2015) 58-67.

Download the Matlab codes:

SO-PLS for multi-block regression

SO-PLS-LDA for multi-block classification