Improving the Competency of Classifiers through Data Generation

  • Herna L. Viktor
  • Iryna Skrypnik
Conference paper


This paper describes a hybrid approach in which sub-symbolic neural networks and symbolic machine learning algorithms are grouped into an ensemble of classifiers. Initially each classifier determines which portion of the data it is most competent in. The competency information is used to generated new data that are used for further training and prediction. The application of this approach in a difficult to learn domain shows an increase in the predictive power, in terms of the accuracy and level of competency of both the ensemble and the component classifiers.


Problem Domain Training Instance Data Generation Process Disjunctive Normal Form Decision Tree Algorithm 
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]
    E Bauer R Kohavi, 1999. An Empirical Comparison of Voting Classification Algorithms: Bagging, Boosting and Variants. Machine Learning, 36, pp.105–139.CrossRefGoogle Scholar
  2. [2]
    L Breiman, 1996. Bagging Predictors. Machine Learning, 24(2), pp.123–140.MathSciNetMATHGoogle Scholar
  3. [3]
    S Cost S. Salzberg, 1993. A Weighted Nearest Neighbor Algorithm for Learning with Symbolic Features. Machine Learning, 10(1), pp. 57–78.Google Scholar
  4. [4]
    P Clark T. Niblett, 1989. The CN2 rule in duction program, Machine learning, pp.261–283.Google Scholar
  5. [5]
    T Dietterich, 1997. Machine Learning Research: Four Current Directions. Artificial Intelligence, 18(4), pp.97–136.Google Scholar
  6. [6]
    J-N Hwang, et al, 1991. Query-based Learning applied to Partially Trained Multilayer Perceptrons, IEEE Transactions on Neural Networks, 2(1), pp.131–136.CrossRefGoogle Scholar
  7. [7]
    R Maclin D Opitz, 1997. An Empirical Evaluation of Bagging and Boosting. Proc. of 14th National Conf. on Artificial Intelligence, AAAI/MIT Press, 1997, pp. 546–55l.Google Scholar
  8. [8]
    JR. Quinlan, 1994. C4.5: Programs for Machine Learning, Morgan Kaufmann, California: USA.Google Scholar
  9. [9]
    R Shapiro Y Freud P Bartlett W Lee, 1997. Boosting the margin: A new explanation of the effectiveness of the voting methods. Proc. of 14th Intern. Conf. on Machine Learning, Morgan Kaufmann, pp. 322–330.Google Scholar
  10. [10]
    SB Thrun et al, 1991. The Monk’s problems: A Performance Comparison of Different Learning Algorithms. Technical Report CMU-CS-91-17. Computer Science Department, Carnegie Mellon University, Pittsburgh: USA.Google Scholar
  11. [11]
    HL Viktor, AP Engelbrecht, I Cloete, 1995. Reduction of Symbolic Rules from Artificial Neural Networks using Sensitivity Analysis, IEEE International Conference on Neural Networks (ICNN’95), Perth: Australia, pp.1788–1793.Google Scholar
  12. [12]
    HL Viktor. 2000. Generating new patterns for information gain and improved neural network learning, The International Joint Conference on Neural Networks (IJCNN-OO), Como: Italy.Google Scholar

Copyright information

© Springer-Verlag Wien 2001

Authors and Affiliations

  • Herna L. Viktor
    • 1
  • Iryna Skrypnik
    • 2
  1. 1.Department of Informatics, School of Information TechnologyUniversity of PretoriaSouth Africa
  2. 2.Department of Computer Science and Information SystemUniversity of JyväskyläFinland

Personalised recommendations