Partial Least Squares Discriminant Analysis (PLS-DA) is a supervised method for classification and regression that combines dimensionality reduction with discriminant analysis. It is used to model high-dimensional data by constructing latent variables (LV) as linear combinations of the original predictor variables. These latent variables explain the relationship with the response variable and aim to find a lower-dimensional representation of the predictors that best discriminate between classes or predict categorical outcomes.
