Advertisement

On changing continuous attributes into ordered discrete attributes

  • J. Catlett
Part 3: Numeric And Statistical Approaches
First Online:
Part of the Lecture Notes in Computer Science book series (LNCS, volume 482)

Abstract

The large real-world datasets now commonly tackled by machine learning algorithms are often described in terms of attributes whose values are real numbers on some continuous interval, rather than being taken from a small number of discrete values. Many algorithms are able to handle continuous attributes, but learning requires far more CPU time than for a corresponding task with discrete attributes. This paper describes how continuous attributes can be converted economically into ordered discrete attributes before being given to the learning system. Experimental results from a wide variety of domains suggest this change of representation does not often result in a significant loss of accuracy (in fact it sometimes significantly improves accuracy), but offers large reductions in learning time, typically more than a factor of 10 in domains with a large number of continuous attributes.

Keywords

Discretisation empirical concept learning induction of decision trees 
This is a preview of subscription content, log in to check access.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Amarel, S. (1968). On the representation of problems of reasoning about action, In D. Michie (Ed.), Machine Intelligence 3, Edinburgh University Press.Google Scholar
  2. Breiman, L., Friedman, J. H., Olshen, R. A., Stone C. J. (1984). Classification and regression trees. Belmont, CA: Wadsworth International Group.Google Scholar
  3. Carter, C., & Catlett, J. (1987). Assessing credit card applications using machine learning. IEEE Expert, Fall 1987, 71–79.Google Scholar
  4. Kononenko, I., Bratko, I., & Roskar, E. (1984). Experiments in automatic learning of medical diagnostic rules, Technical Report, Jozef Stefan Institute, Ljubljana.Google Scholar
  5. Michie, D. (1987). Current developments in expert systems. In J. R. Quinlan, (Ed.), Applications of Expert Systems. Maidenhead: Addison Wesley.Google Scholar
  6. Michalski, R., Mozetic, T., Hong, J., Lavrac, N. (1986). The multi-purpose incremental learning system AQ15 and its testing application to three medical domains Proceedings of AAAI-86, Morgan Kaufmann.Google Scholar
  7. Oates, J., Cellar, B., Bernstein, L., Bailey, B. P., Freedman, S. B. (1989). Real-time detection of ischemic ECG changes using quasi-orthogonal leads and artificial intelligence, Proceedings, IEEE Computers in Cardiology Conference, 1989, IEEE Computer Society.Google Scholar
  8. Quinlan, J. R. (1979). Discovering rules by induction from large numbers of examples: a case study. In D. Michie (Ed.), Expert systems in the micro-electronic age. Edinburgh University Press.Google Scholar
  9. Quinlan, J. R. (1983). Learning efficient classification procedures and their application to chess endgames (p. 469). In R. S. Michalski, J. R. Carbonell, T. M. Mitchell (Eds.), Machine learning: an Artificial Intelligence approach (pp. 463–82). Los Altos, CA: Morgan Kaufmann.Google Scholar
  10. Quinlan, J. R. (1986). Induction of decision trees. Machine Learning, 1,1.Google Scholar
  11. Quinlan, J. R., Compton, P.J., Horn, K.A. & Lazarus, L. (1988). Inductive knowledge acquisition: a case study In J. Quinlan (Ed.), Applications of expert systems, Maidenhead: Addison-WesleyGoogle Scholar
  12. Quinlan, J. R. (1987). Simplifying decision trees. International Journal of Man-machine Studies, 27 (pp. 221–234).Google Scholar
  13. Quinlan, J. R. (1987b). Decision trees as probabilistic classifiers, Proceedings of the fourth international conference on machine learning, (pp. 31–37) Morgan Kaufmann.Google Scholar
  14. Quinlan, J. R. (1989). Unknown attribute values in induction Proceedings of the sixth international conference on machine learning, (pp. 164–168) Morgan Kaufmann.Google Scholar
  15. Rendell, L. (1989). Comparing systems and analysing functions to improve constructive induction Proceedings of the sixth international conference on machine learning (pp. 461–464) Morgan Kaufmann.Google Scholar
  16. Sejnowski, T. J., & Rosenberg, C. R., (1987). Parallel networks that learn to pronounce English text Complex Systems 1. (pp. 426–429).Google Scholar
  17. Subramanian, D. (1989). Representational issues in machine learning Proceedings of the sixth international conference on machine learning (pp. 426–429) Morgan KaufmannGoogle Scholar
  18. Utgoff, P. & Heitman, P.S. (1988). Learning and generalizing move selection preferences Proceedings of the AAAI symposium on computer game playing pp. 36–40 (original not seen).Google Scholar
  19. Utgoff, P. (1989). ID5: an incremental ID3 Proceedings of the fifth international conference on machine learning (pp. 107–120) Morgan Kaufmann.Google Scholar
  20. Wilson, S. W. (1987). Classifier systems and the animat problem Machine Learning, 2,4.Google Scholar
  21. Wirth, J., & Catlett, J. (1988). Costs and benefits of windowing in ID3 Proceedings of the fifth international conference on machine learning (pp. 87–99) Morgan Kaufmann.Google Scholar
  22. Wong, A.K.C., Chiu, D.K.Y., (1987). Synthesizing statistical knowledge from incomplete mixed-mode data, IEEE Trans. Pattern Analysis and Machine Intelligence, November 1987, Vol PAMI-9, No. 6, pp. 796–805.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • J. Catlett
    • 1
  1. 1.Basser Department of Computer ScienceUniversity of SydneyAustralia

Personalised recommendations

Cite paper

Buy options