Partial least squares regression

Introduction
1. Partial least squares regression (PLSR) is a method that combines principal component analysis and multiple regression. PLSR performs a descriptor dimension reduction procedure and constructs a set of components that accounts for as much as possible of the total descriptors variance in the dataset. This helps to avoid multicollinearity and overfitting of the model.
2. The PLSR method is more suitable for modeling small training sets using large sets of descriptors. On ViNAS, the PLSR algorithm is implemented using scikit-learn 0.24.1.