Improving Numerical Prediction with Qualitative Constraints

  • Dorian Šuc
  • Ivan Bratko
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2837)


The usual numerical learning methods, that are primarily concerned with finding a good numerical fit to the data, often make predictions that do not correspond to the qualitative mechanisms in the domain of modelling or a domain expert’s intuition. Consistency of numerical predictions with a given qualitative model is helpful when a numerical model is used for explanation of phenomena in the modelled domain, but can also considerably improve numerical accuracy. In this paper we present a novel approach to numerical machine learning called Qfilter. Qfilter is a numerical regression method that can take into account qualitative background knowledge to give qualitatively faithful numerical prediction. The results on a set of domains including population dynamics show considerable prediction accuracy improvements compared to the usual numerical learners. As qualitative domain knowledge is often available in practice, Qfilter’s ability to exploit such knowledge should be beneficial in many applications.


Mean Square Error Numerical Prediction Qualitative Model Zooplankton Population Monotonicity Constraint 
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.


  1. 1.
    Šuc, D., Vladušič, D., Bratko, I.: Qualitatively faithful quantitative prediction. In: Proceedings of the 18th International Joint Conference on Artificial Intelligence, Acapulco, Mexico (August 2003)Google Scholar
  2. 2.
    Quinlan, J.: Learning with continuous classes. In: Proc. of the 5th Australian Joint Conference on Artificial Intelligence, pp. 343–348. World Scientific, Singapore (1992)Google Scholar
  3. 3.
    Atkeson, C., Moore, A., Schaal, S.: Locally weighted learning. Artificial Intelligence Review 11, 11–73 (1997)CrossRefGoogle Scholar
  4. 4.
    Witten, I., Frank, E.: Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations, vol. ch. 8, pp. 265–320. Morgan Kaufmann, San Francisco (2000)Google Scholar
  5. 5.
    Šuc, D.: Machine Reconstruction of Human Control Strategies. PhD thesis, Faculty of Computer and Information Sc., University of Ljubljana, Slovenia (2001)Google Scholar
  6. 6.
    Šuc, D., Bratko, I.: Induction of qualitative trees. In: De Raedt, L., Flach, P. (eds.) Proc. of the 12th European Conf. on Machine Learning, pp. 442–453. Springer, Heidelberg (2001)Google Scholar
  7. 7.
    Šuc, D., Bratko, I.: Qualitative reverse engineering. In: Sammut, C., Hoffmann, A. (eds.) Proc. of the 19th International Conf. on Machine Learning, pp. 610–617. Morgan Kaufmann, San Francisco (2002)Google Scholar
  8. 8.
    Forbus, K.: Qualitative process theory. Artificial Intelligence 24, 85–168 (1984)CrossRefGoogle Scholar
  9. 9.
    Kuipers, B.: Qualitative simulation. Artificial Intelligence 29, 289–338 (1986)zbMATHCrossRefMathSciNetGoogle Scholar
  10. 10.
    The MathWorks, I.: Matlab software (2003),
  11. 11.
    Coleman, T.F., Li, Y.: A reflective Newton method for minimizing a quadratic function subject to bounds on some of the variables. SIAM Journal on Optimization 6, 1040–1058 (1996)zbMATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Gill, P.E., Murray, W., Wright, M.H.: Quadratic programming. In: Practical Optimization, pp. 177–184. Academic Press, London (1981)Google Scholar
  13. 13.
    Todorovski, L., Džeroski, S.: Using domain knowledge on population dynamics modeling for equation discovery. In: Proceedings of the 12th European Conference on Machine Learning, pp. 478–490. Springer, Heidelberg (2001)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Dorian Šuc
    • 1
  • Ivan Bratko
    • 1
  1. 1.Faculty of Computer and Information ScienceUniversity of LjubljanaLjubljanaSlovenia

Personalised recommendations