JointEntropy/awesome-ml-pu-learning

This is a curated collection of research papers and software implementations focused on Positive Unlabeled (PU) learning, a specialized machine learning technique. It helps data scientists and machine learning engineers build more accurate classification models when they have plenty of examples for one category (like 'known spam') but can't definitively label the other category ('not spam' vs. 'unknown'). The resource provides various methods to train models using data that is either positively labeled or completely unlabeled, improving model performance in challenging real-world scenarios.