PU Learning Tutorial

An introductory tutorial to the "Learning from Positive and Unlabeled Data" field.

This is an introductory tutorial to the “Learning from Positive and Unlabeled Data” field, which consists of 6 parts:

  1. PU Learning and its sources
  2. PU Learning definitions
  3. Assumptions to enable PU Learning
  4. Two-step techniques
  5. Biased learning
  6. Incorporation of the labeling mechanism

The content of the tutorial is largely based on our survey paper: Learning from positive and unlabeled data: a survey, which is freely available on arXiv.