Performance for the whole training set, 1^st degree classifiable and 1^st degree unclassifiable peptides employing the simplified machine learning approach with human-understandable attributes (ML-simple) for prediction of IVIG binding^*.

*Comparison of AUC and accuracy when the classifier was 10-fold cross-validated on all the peptides in the original training set, or on peptides that the first classifier classified correctly (1st degree classifiable) or incorrectly (1st degree unclassifiable), respectively. All classifiers used the interpretable attributes and logistic regression.