Great list. I would also include GLMSELECT in the list (stepwise) for ML type applications where you don't know what are the best features to use in a predictive model. For example, a system state may be described by hundreds of features and you want to predict the future state of one of those features given the past state of the total system. GLMSELECT can help you select the best features to use for a regression model. In addition, you can use training, test and validation samples to minimize overfitting. ... View more