A Genetic Inspired Optimization for ECOC

  • Miguel Ángel Bautista
  • Sergio Escalera
  • Xavier Baró
  • Oriol Pujol
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7626)

Abstract

In this work, we propose a novel Genetic Inspired Error Correcting Output Codes (ECOC) Optimization, which looks for an efficient problem-dependent encoding of the multi-class task with high generalization performance. This optimization procedure is based on novel ECOC-Compliant crossover, mutation, and extension operators, which guide the optimization process to promising regions of the search space. The results on several public datasets show significant performance improvements as compared to state-of-the-art ECOC strategies.

Keywords

Error-Correcting Output Codes Genetic Optimization Ensemble learning 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Allwein, E., Schapire, R., Singer, Y.: Reducing multiclass to binary: A unifying approach for margin classifiers. JMLR 1, 113–141 (2002)MathSciNetGoogle Scholar
  2. 2.
    Ángel Bautista, M., Escalera, S., Baró, X., Radeva, P., Vitriá, J., Pujol, O.: Minimal design of error-correcting output codes. PRL 33(6), 693–702 (2012)CrossRefGoogle Scholar
  3. 3.
    Demsar, J.: Statistical comparisons of classifiers over multiple data sets. JMLR 7, 1–30 (2006)MathSciNetMATHGoogle Scholar
  4. 4.
    Escalera, S., Pujol, O., Radeva, P.: On the decoding process in ternary error-correcting output codes. TPAMI 99(1) (2009)Google Scholar
  5. 5.
    Garcia-Pedrajas, N., Fyfe, C.: Evolving output codes for multiclass problems. TEC 12(1), 93–106 (2008)Google Scholar
  6. 6.
    Lorena, A.C., Carvalho, A.C.P.L.F.: Evolutionary design of multiclass support vector machines. JIFS 18, 445–454 (2007)MATHGoogle Scholar
  7. 7.
    Pujol, O., Radeva, P., Vitrià, J.: Discriminant ECOC: A heuristic method for application dependent design of ecoc. TPAMI 28, 1001–1007 (2006)CrossRefGoogle Scholar
  8. 8.
    Escalera, S., Pujol, O., Radeva, P.: Boosted landmarks and forest ECOC: Framework to detect and classify objects in clutter scenes. PRL 28(13), 1759–1768 (2007)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Miguel Ángel Bautista
    • 1
    • 2
  • Sergio Escalera
    • 1
    • 2
  • Xavier Baró
    • 2
    • 3
  • Oriol Pujol
    • 1
    • 2
  1. 1.Dept. Matemàtica Aplicada i AnàlisiUniversitat de BarcelonaBarcelonaSpain
  2. 2.Centre de Visió per ComputadorCampus UAB, Edifici OBarcelonaSpain
  3. 3.EIMS, Universitat Oberta de CatalunyaBarcelonaSpain

Personalised recommendations