On the unknown attribute values in learning from examples

  • Jerzy W. Grzymala-Busse
Communications Learning and Adaptive Systems
Part of the Lecture Notes in Computer Science book series (LNCS, volume 542)


In machine learning many real-life applications data are characterized by attributes with unknown values. This paper shows that the existing approaches to learning from such examples are not sufficient. A new method is suggested, which transforms the original decision table with unknown values into a new decision table in which every attribute value is known. Such a new table, in general, is inconsistent. This problem is solved by a technique of learning from inconsistent examples, based on rough set theory. Thus, two sets of rules: certain and possible are induced. Certain rules are categorical, while possible rules are supported by existing data, although conflicting data may exist as well. The presented approach may be combined with any other approach to uncertainty when processing of possible rules is concerned.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    T. Arciszewski, M. Mustafa, and W. Ziarko. A methodology of design knowledge acquisition for use in learning expert systems. Int. J. Man-Machine Studies 27, 1987, 23–32.Google Scholar
  2. [2]
    J. Catlett. Induction using the Shafer representation. Technical Report. Basser Department of Computer Science, University of Sydney, Australia, 1985.Google Scholar
  3. [3]
    C.-C. Chan and J. W. Grzymala-Busse. Rough-set boundaries as a tool for learning rules from examples. Proc. ISMIS-89, 4th Int. Symposium on Methodologies for Intelligent Systems, 1989, 281–288.Google Scholar
  4. [4]
    P. Clark, T. Niblett. The CN2 induction algorithm. Machine Learning 3, 1989, 261–283.Google Scholar
  5. [5]
    T. G. Dietterich, R. S. Michalski. A comparative review of selected methods for learning from examples. In Machine Learning. An Artificial Intelligence Approach, ed. R. S. Michalski, J. G. Carbonell, T. M. Mitchell, Morgan Kauffman, 1983, 41–81.Google Scholar
  6. [6]
    J. H. Friedman. A recursive partitioning decision rule for nonparametric classification. IEEE Transactions on Computers, 1977, 404–408.Google Scholar
  7. [7]
    J. W. Grzymala-Busse. Knowledge acquisition under uncertainty—a rough set approach. Journal of Intelligent & Robotic Systems 1, 1988, 3–16.Google Scholar
  8. [8]
    I. Knonenko, I. Bratko, E. Roskar. Experiments in automatic learning of medical diagnostic rules. Technical Report, Jozef Stefan Institute, Ljubljana, Yugoslavia, 1984.Google Scholar
  9. [9]
    D. Maier. The Theory of Relational Databases, Computer Science Press, 1983.Google Scholar
  10. [10]
    M. V. Manago, Y. Kodratoff. Noise and knowledge acquisition. Proc. IJCAI 1987, 348–354.Google Scholar
  11. [11]
    R. S. Michalski. A theory and methodology of inductive learning. In Machine Learning. An Artificial Intelligence Approach, ed. R. S. Michalski, J. G. Carbonell, T. M. Mitchell, Morgan Kauffman, 1983, 83–134.Google Scholar
  12. [12]
    Z. Pawlak. Rough sets. Int. J. Computer and Information Sci. 11, 1982, 341–356.Google Scholar
  13. [13]
    J. R. Quinlan. Induction of decision trees. Machine Learning 1, 1986, 81–106.Google Scholar
  14. [14]
    J. R. Quinlan. Decision trees as probabilistic classifiers. Proc. 4th Int. Workshop on Machine Learning 1987, 31–37.Google Scholar
  15. [15]
    J. R. Quinlan. Unknown attribute values in induction. Proc. 6th Int. Workshop on Machine Learning, 1989, 164–168.Google Scholar
  16. [16]
    J. R. Quinlan. Probabilistic decision trees. In Machine Learning. An Artificial Intelligence Approach, vol III, ed. Y. Kodratoff and R. S. Michalski, 1990, 140–152.Google Scholar
  17. [17]
    R. Yasdi and W. Ziarko. An expert system for conceptual schema design: A machine learning approach. Int. J. Man-Machine Studies 29, 1988, 351–376.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Jerzy W. Grzymala-Busse
    • 1
  1. 1.Department of Computer ScienceUniversity of KansasLawrence

Personalised recommendations