The goal of this workflow is to classify the car makers based on the cars they produce. The target class is the "make" column. The calssification model could a decision tree, a pNN, or a manual set of rules: this is the user choice. The choice for the classification model is made in the Value Selection Widget node. Depending on which predictive model has been chosen, the corresponding branch of the CASE block is activated and the model is trained and applied. Predictions are collected in the CASE Switch Data (End) node and performances on the training set are evaluated at the end with a Scorer node. The trick is to produce predictions in an output column named the same on all the branches. The selected model is collected with a CASE Switch Model (End) node and written to file.