Positive-Unlabelled (PU) learning is a semi-supervised binary classification technique that is trained on dataset where only positive instances are labelled, and remainder of instances are unlabelled/unknown.

Supported by