Advertisement

Learning Fuzzy Models and Potential Outliers

  • M. R. Berthold
Conference paper
Part of the International Centre for Mechanical Sciences book series (CISM, volume 408)

Abstract

Outliers or distorted attributes very often severely interfere with data analysis algorithms that try to extract few meaningful rules. Most methods to deal with outliers try to completely ignore them. This can be potentially harmful since the very outlier that was ignored might have described a rare but still extremely interesting phenomena. In this paper we describe an approach that tries to build an interpretable model while still maintaining all the information in the data. This is achieved through a two stage process. A first phase builds an outlier-model for data points of low relevance, followed by a second stage which uses this model as filter and generates a simpler model, describing only examples with higher relevance, thus representing a more general concept. The outlier-model on the other hand may point out potential areas of interest to the user. Preliminary experiments indicate that the two models in fact have lower complexity and sometimes even offer superior performance.

Keywords

Membership Function Fuzzy Rule Fuzzy Model Potential Outlier Fuzzy Graph 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    C. Apte, S. Hong, J. Hosking, J. Lepre, E. Pednault, and B. K. Rosen. Decomposition of heterogeneous classification problems. Intelligent Data Analysis, 2 (2), 1998. ( http://www.elsevier.nl/locate/ida ).
  2. 2.
    M. R. Berthold and K.-P. Huber. Constructing fuzzy graphs from examples. Intelligent Data Analysis, 3 (1), 1999. ( http://www.elsevier.nl/locate/ida).Google Scholar
  3. 3.
    V. Cherkassky and F. Mulier. Learning from Data. John Wiley and Sons Inc., 1998.Google Scholar
  4. 4.
    S. K. Halamuge and M. Glesner. FuNe Deluxe: A group of fuzzy-neural methods for complex data analysis problems. In Proceedings of the EUFIT’95, Aug. 1995.Google Scholar
  5. 5.
    D. Hand, J. Kok, and M. Berthold, editors. Advances in Intelligent Data Analysis. LNCS. Springer Verlag, 1999.CrossRefGoogle Scholar
  6. 6.
    C. M. Higgins and R. M. Goodman. Learning fuzzy rule-based neural networks for control. In Advances in Neural Information Processing Systems 5, pages 350–357, California, 1993. Morgan Kaufmann.Google Scholar
  7. 7.
    K.-P. Huber and M. R. Berthold. Building precise classifiers with automatic rule extraction. In IEEE International Conference on Neural Networks, 3, pages 12631268, 1995.Google Scholar
  8. 8.
    D. Michie, D. J. Spiegelhalter, and C. C. Taylor, editors. Machine Learning, Neural and Statistical Classification. Ellis Horwood Limited, 1994.Google Scholar
  9. 9.
    R. Silipo and M. R. Berthold. Discriminative power of input features. In [5], 1999.Google Scholar
  10. 10.
    P. K. Simpson. Fuzzy min-max neural networks — part 1: Classification. IEEE Transactions on Neural Networks, 3 (5): 776–786, Sept. 1992.CrossRefGoogle Scholar
  11. 11.
    P. K. Simpson. Fuzzy min-max neural networks — part 2: Clustering. IEEE Transactions on Fuzzy Systems, 1 (1): 32–45, Jan. 1993.CrossRefGoogle Scholar
  12. 12.
    L.-X. Wang and J. M. Mendel. Generating rules by learning from examples. In International Symposium on Intelligent Control, pages 263–268. IEEE, 1991.Google Scholar
  13. 13.
    L.-X. Wang and J. M. Mendel. Generating fuzzy rules by learning from examples. IEEE Transactions on Systems, Man, and Cybernetics, 22 (6): 1313–1427, 1992.MathSciNetGoogle Scholar
  14. 14.
    L. A. Zadeh. Soft computing and fuzzy logic. IEEE Software, pages 48–56, Nov. 1994.Google Scholar
  15. 15.
    L. A. Zadeh. Fuzzy logic and the calculi of fuzzy rules and fuzzy graphs: A precis. Multi. Val. Logic, 1: 1–38, 1996.MathSciNetMATHGoogle Scholar

Copyright information

© Springer-Verlag Wien 2000

Authors and Affiliations

  • M. R. Berthold
    • 1
  1. 1.University of California at BerkeleyBerkeleyUSA

Personalised recommendations