Skip to content

C4.5: Programs for Machine Learning

Authors: J. Ross Quinlan

Published: 1993 (Other)

Source: Morgan Kaufmann

Algorithm: C4.5

Summary

Extends the ID3 algorithm into C4.5, adding gain ratio (to correct ID3's bias toward high-cardinality attributes), cost-sensitive learning, handling of continuous attributes, pruning by estimated error rate, and direct handling of missing values. C4.5 was historically one of the most used decision-tree systems and the basis for Weka's J48 classifier.

Abstract

Primary

Paper dl.acm.org

Tags

  • Decision trees

  • C4.5

  • Classification

  • Machine learning

  • Gain ratio

  • Pruning