Problem Definition

Kristen Jaskie, Andreas Spanias

Research output: Chapter in Book/Report/Conference proceedingChapter


In this chapter, we present the Positive and Unlabeled (PU) learning problem in detail. In addition to notation and a formal statement of the problem, we explore fundamental factors that influence PU models such as how the source data for the model was acquired and what assumptions are made about that data. We end the chapter with a brief description of two closely related learning problems, one class classification (OCC) and noisy learning.

Original languageEnglish (US)
Title of host publicationSynthesis Lectures on Artificial Intelligence and Machine Learning
PublisherSpringer Nature
Number of pages22
StatePublished - 2022

Publication series

NameSynthesis Lectures on Artificial Intelligence and Machine Learning
ISSN (Print)1939-4608
ISSN (Electronic)1939-4616

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this