pu-learn

A library for learning classifiers only with positive and unlabeled data. Adheres to the scikit-learn API

Models Implemented (Reference)

• RUE: Reweighted Unlabeled Examples (Learning Classifiers from Only Positive and Unlabeled Data)
• I-EM: Iterative Expectation Maximization (Partially Supervised Classification of Text Documents)
• S-EM: Spy-Iterative Expectation Maximization (Partially Supervised Classification of Text Documents)