Fig. 2

ROC curves from the random forest classifier for the training (orange) and validation (blue) sets. The star represents the model whose performance is referred to in the text (sensitivity of 95% on the training set)