A positive and unlabeled learning (PUL) problem occurs when a machine learning set of training data has only a few positive labeled items and many unlabeled items. PUL problems often occur with ...