Machine Learning

, Volume 24, Issue 1, pp 41–47 | Cite as

Technical note: Some properties of splitting criteria

  • Leo Breiman


Various criteria have been proposed for deciding which split is best at a given node of a binary classification tree. Consider the question: given a goodness-of-split criterion and the class populations of the instances at a node, what distribution of the instances between the two children nodes maximizes the goodness-of-split criterion? The answers reveal an interesting distinction between the gini and entropy criterion.


Trees Classification Splits 


  1. Breiman, L., Friedman, J., Olshen, R., & Stone, C. (1984). “Classification and Regression Trees,” Wadsworth.Google Scholar
  2. Buntine, W. & Niblett, T. (1992). “A further comparison of splitting rules for decision tree induction,” Machine Learning 8, 75–85.Google Scholar
  3. Fayyad, U.M. (1991). “On the induction of decision trees for multiple concept learning,” Ph.D Thesis, Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, Michigan.Google Scholar
  4. Fayyad, U.M. & Irani, R.B. (1990). “What should be minimized in a decision tree”? Proc. 8th National Conf. on AI, AAAI-90, 749–754, MIT Press.Google Scholar
  5. Fayyad, V.M. & Irani, R.B. (1992). “The attribute selection problem in decision tree generation,” Proc. 10th National Conf. on AI, AAAI-92, 104–110, MIT Press.Google Scholar
  6. Fayyad, U.M. & Irani, R.B. (1993). “Multi-interval discretization of continuous valued attribute for classification learning,” Proc. 13th International Joint Conf. on AI, 1022–1027, Morgan Kaufmann.Google Scholar
  7. Taylor, P.C. & Silverman, B.W. (1993). “Block diagrams and splitting criteria for classification trees,” Statistics and Computing, V 3, p. 147–161.Google Scholar

Copyright information

© Kluwer Academic Publishers 1996

Authors and Affiliations

  • Leo Breiman
    • 1
  1. 1.Statistics DepartmentUniversity of CaliforniaBerkeley

Personalised recommendations