Designing Fusers on the Basis of Discriminants – Evolutionary and Neural Methods of Training

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6076)


The combining approach to classification is nowadays one of the most promising directions in pattern recognition. There are many methods of decision-making that can be used by an ensemble of classifiers. The most popular methods have their origins in voting, where the decision of a common classifier is a combination of individual classifiers’ outputs, i.e. class numbers or values of discriminants. This work focuses on the problem of fuser design. We propose to train a fusion block by algorithms that have their origin in neural and evolutionary approaches. As we have shown in previous works, we can produce better combining classifiers than Oracle can. Presented results of experiments confirm our previous observations.


combining classifiers evolutionary algorithms neural networks fuser design 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Kuncheva, L.I., Bezdek, J.C., Duin, R.P.W.: Decision templates for multiple classifier fusion: an experimental comparison. Pattern Recognition 34, 299–314 (2001)CrossRefzbMATHGoogle Scholar
  2. 2.
    Kuncheva, L.I.: Combining pattern classifiers: Methods and algorithms. Wiley, Chichester (2004)CrossRefzbMATHGoogle Scholar
  3. 3.
    Biggio, B., Fumera, G., Roli, F.: Bayesian Analysis of Linear Combiners. In: Haindl, M., Kittler, J., Roli, F. (eds.) MCS 2007. LNCS, vol. 4472, pp. 292–301. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  4. 4.
    Duin, R.P.W., et al.: PRTools4, A Matlab Toolbox for Pattern Recognition. Delft University of Technology, The Netherlands (2004)Google Scholar
  5. 5.
    Woods, K., Kegelmeyer, W.P.: Combination of multiple classifiers using local accuracy estimates. IEEE Transactions on PAMI 19(4), 405–410 (1997)CrossRefGoogle Scholar
  6. 6.
    Woźniak, M., Jackowski, K.: Some remarks on chosen methods of classifier fusion based on weighted voting. In: Corchado, E., Wu, X., Oja, E., Herrero, Á., Baruque, B. (eds.) HAIS 2009. LNCS, vol. 5572, pp. 541–548. Springer, Heidelberg (2009)CrossRefGoogle Scholar
  7. 7.
    Michalewicz, Z.: Genetics Algorithms + Data Structures = Evolutions Programs. Springer, Berlin (1996)CrossRefzbMATHGoogle Scholar
  8. 8.
    Asuncion, A., Newman, D.J.: UCI ML Repository, Irvine, CA: University of California, School of Information and Computer Science (2007),
  9. 9.
    Zmyślony, M., Woźniak, M.: Influence of fusion methods on quality of classification. Advanced Simulation of Systems, 117–120 (2010)Google Scholar
  10. 10.
    Kuncheva, L.I., Bezdek, J.C., Duin, R.P.W.: Decision templates for multiple classifier fusion: an experimental comparison. Pattern Recognition 34, 299–314 (2001)CrossRefzbMATHGoogle Scholar
  11. 11.
    Giacinto, G., Roli, F., Fumera, G.: Design of Effective Multiple Classifier Systems by Clustering of Classifiers. In: Proceedings of the 15th International Conference on Pattern Recognition (ICPR 2000), vol. 2, p. 2160 (2000)Google Scholar
  12. 12.
    Marcialis, G.L., Roli, F.: Fusion of Face Recognition Algorithms for Video-Based Surveillance Systems. In: Foresti, G.L., Regazzoni, C., Varshney, P. (eds.) Multisensor Surveillance Systems: The Fusion Perspective. Kluwer Academic Publishers, Dordrecht (2003)Google Scholar
  13. 13.
    Wolpert, D.H.: The supervised learning no-free-lunch theorems. In: Proceedings of the 6th Online World Conference on Soft Computing in Industrial Applications (2001)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  1. 1.Department of Systems and Computer NetworksWroclaw University of TechnologyWroclawPoland

Personalised recommendations