Efficiently Explaining Decisions of Probabilistic RBF Classification Networks

  • Marko Robnik-Šikonja
  • Aristidis Likas
  • Constantinos Constantinopoulos
  • Igor Kononenko
  • Erik Štrumbelj
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6593)


For many important practical applications model transparency is an important requirement. A probabilistic radial basis function (PRBF) network is an effective non-linear classifier, but similarly to most other neural network models it is not straightforward to obtain explanations for its decisions. Recently two general methods for explaining of a model’s decisions for individual instances have been introduced which are based on the decomposition of a model’s prediction into contributions of each attribute. By exploiting the marginalization property of the Gaussian distribution, we show that PRBF is especially suitable for these explanation techniques. By explaining the PRBF’s decisions for new unlabeled cases we demonstrate resulting methods and accompany presentation with visualization technique that works both for single instances as well as for the attributes and their values, thus providing a valuable tool for inspection of the otherwise opaque models.


classification explanation model explanation comprehensibility probabilistic RBF networks model visualization game theory 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Bishop, C.M.: Neural Networks for Pattern Recognition. Oxford University Press, Oxford (1995)zbMATHGoogle Scholar
  2. 2.
    Constantinopoulos, C., Likas, A.: An incremental training method for the probabilistic RBF network. IEEE Trans. Neural Networks 17(4), 966–974 (2006)CrossRefzbMATHGoogle Scholar
  3. 3.
    Jacobsson, H.: Rule extraction from recurrent neural networks: A taxonomy and review. Neural Computation 17(6), 1223–1263 (2005)MathSciNetCrossRefzbMATHGoogle Scholar
  4. 4.
    McLachlan, G., Peel, D.: Finite Mixture Models. John Wiley & Sons, Chichester (2000)CrossRefzbMATHGoogle Scholar
  5. 5.
    Možina, M., Demšar, J., Kattan, M.W., Zupan, B.: Nomograms for visualization of naive bayesian classifier. In: Boulicaut, J.-F., Esposito, F., Giannotti, F., Pedreschi, D. (eds.) PKDD 2004. LNCS (LNAI), vol. 3202, pp. 337–348. Springer, Heidelberg (2004)Google Scholar
  6. 6.
    Robnik Šikonja, M., Kononenko, I.: Explaining classifications for individual instances. IEEE Transactions on Knowledge and Data Engineering 20(5), 589–600 (2008)CrossRefGoogle Scholar
  7. 7.
    Setiono, R., Liu, H.: Understanding neural networks via rule extraction. In: Proceedings of IJCAI 1995, pp. 480–487 (1995)Google Scholar
  8. 8.
    Shapley, L.S.: A value for n-person games. In: Contributions to the Theory of Games, vol. II. Princeton University Press, Princeton (1953)Google Scholar
  9. 9.
    Titsias, M.K., Likas, A.: Shared kernel models for class conditional density estimation. IEEE Trans. Neural Networks 12(5), 987–997 (2001)CrossRefGoogle Scholar
  10. 10.
    Titsias, M.K., Likas, A.: Class conditional density estimation using mixtures with constrained component sharing. IEEE Trans. Pattern Anal. and Machine Intell. 25(7), 924–928 (2003)CrossRefGoogle Scholar
  11. 11.
    Towell, G.G., Shavlik, J.W.: Extracting refined rules from knowledge-based neural networks. Machine Learning 13(1), 71–101 (1993)Google Scholar
  12. 12.
    Štrumbelj, E., Kononenko, I.: An Efficient Explanation of Individual Classifications using Game Theory. Journal of Machine Learning Research 11, 1–18 (2010)MathSciNetzbMATHGoogle Scholar
  13. 13.
    Štrumbelj, E., Kononenko, I., Robnik-Šikonja, M.: Explaining instance classifications with interactions of subsets of feature values. Data & Knowledge Engineering 68(10), 886–904 (2009)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Marko Robnik-Šikonja
    • 1
  • Aristidis Likas
    • 2
  • Constantinos Constantinopoulos
    • 3
  • Igor Kononenko
    • 1
  • Erik Štrumbelj
    • 1
  1. 1.Faculty of Computer and Information ScienceUniversity of LjubljanaLjubljanaSlovenia
  2. 2.Department of Computer ScienceUniversity of IoanninaIoanninaGreece
  3. 3.Centre d’InnovacióBarcelona MediaBarcelonaSpain

Personalised recommendations