Penalization, functionalization, sparsity, robustness... everything can be added (in any combination) to PLS regression, making it a more versatile tool for particular task. Here, we have new two approaches: sparse and robust (with application to classification)  and functional with penalization .
- I. Hoffmann, P. Filzmoser, S. Serneels, and K. Varmuza, "Sparse and robust PLS for binary classification", Journal of Chemometrics, vol. 30, pp. 153-162, 2016. http://dx.doi.org/10.1002/cem.2775
- A. Aguilera, M. Aguilera-Morillo, and C. Preda, "Penalized versions of functional PLS regression", Chemometrics and Intelligent Laboratory Systems, vol. 154, pp. 80-92, 2016. http://dx.doi.org/10.1016/j.chemolab.2016.03.013