On the Lower Boundaries in Learning Rules from Examples

  • Chien-Chung Chan
  • Jerzy W. Grzymala-Busse
Part of the Studies in Fuzziness and Soft Computing book series (STUDFUZZ, volume 13)


The paper studies multi-concept learning from inconsistent examples. The main assumption is that knowledge is acquired in the form of production rules, while induced rules represent minimal discriminant description. The paper presents an approach to the elimination of irrelevant attributes in the concept description, or, to be more specific, a new method for determining coverings, the minimal sets of relevant attributes. This method is based on a new tool created within rough set theory. In the presented approach to learning rules from examples, methods of rough set theory are used twice: first, in defining lower boundaries, and then, to deal with inconsistencies. Inconsistent representation of examples is a kind of uncertainty in input data, resulting from gathering information from inconsistent experts or lack of sufficient number of attributes to describe input data.


Production Rule Learn Rule Approximation Space Irrelevant Attribute Conceptual Variable 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    C.-C. Chan, J. W. Grzymala-Busse. Rough-set boundaries as a tool for learning rules from examples. Proc. 4th Int. Symp. on Methodologies for Intelligent Systems, Oct. 12–14, 1989, Charlotte, NC,281–288Google Scholar
  2. [2]
    J. S. Dean, J. W. Grzymala-Busse. An overview of the learning from examples module LEM1. Report TR-88–2, Department of Computer Science, University of Kansas, 1988Google Scholar
  3. [3]
    D. H. Fisher, J. C. Schlimmer. Concept simplification and prediction accuracy. Proc. 5th Conf on Machine Learning, University ofMichigan, Ann Arbor, MI, June 12–14, 1988, 22–28Google Scholar
  4. [4]
    J. W. Grzymala-Busse. Knowledge acquisition under uncertainty: a rough set approach. Journal of Intelligent and Robotic Systems 1, 1988, 3–16MathSciNetCrossRefGoogle Scholar
  5. [5]
    J. W. Grzymala-Busse. On the learning of minimal discriminant rules from examples. Report TR-89–3, Department of Computer Science, University of Kansas, 1989Google Scholar
  6. [6]
    J. W. Grzymala-Busse. Learning minimal discriminant rules from examples with inconsistencies - a rough set approach. Report TR-89–4, Department of Computer Science, University of Kansas, 1989Google Scholar
  7. [7]
    J. W. Grzymala-Busse, D. J. Sikora. LERS1 — a system for learning from examples based on rough set theory. Report TR-88–3, Department of Computer Science, University ofKansas, 1988Google Scholar
  8. [8]
    R. S. Michalski. A theory and methodology of inductive learning. In Machine Learning. R S. Michalski, J. G. Carbonell, T. M. Mitchell (eds), Tioga Publ. Co., 1983, 83–134Google Scholar
  9. [9]
    Z. Pawlak. Rough sets. Basic notions. Institute Comp. Sci. Polish Acad. Sci. Rep. #431, Warsaw, 1981. Also in Int. J. Computer and Information Sci. 11, 1982, 341–356Google Scholar
  10. [10]
    Z. Pawlak. Rough classification. Int. J. Man Machine Studies 20, 1984, 469–483MATHCrossRefGoogle Scholar
  11. [11]
    J. D. Ullman. Principles of Database Systems. Computer Science Press, 1982Google Scholar
  12. [12]
    A. Wasilewska. Definable sets in knowledge representation systems. Bull. Polish Acad. Sci., Math. 35, 1987, 629–635MathSciNetMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1998

Authors and Affiliations

  • Chien-Chung Chan
    • 1
  • Jerzy W. Grzymala-Busse
    • 2
  1. 1.Department of Mathematical SciencesUniversity of AkronAkronUSA
  2. 2.Department of Electrical Engineering and Computer ScienceUniversity of KansasLawrenceUSA

Personalised recommendations