ICANN 98 pp 171-176 | Cite as

An Analysis of Convergence in Generalized LVQ

  • Atsushi Sato
  • Keiji Yamada
Part of the Perspectives in Neural Computing book series (PERSPECT.NEURAL)


We have proposed a new formulation of Learning Vector Quantization (LVQ) called “Generalized LVQ” based on Minimum Classification Error (MCE). In this paper, we attempt to clarify the convergence property of reference vectors in our formulation. We discuss the equilibrium in a dynamical system for two-class classification, and prove that equilibrium states exist in our formulation, while they do not exist in LVQ2.1 or Juan & Katagiri’s formulation based on MCE.


Equilibrium Point Discriminant Function Convergence Property Learning Rule Convergence Condition 
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]
    T. Kohonen, “Self-Organizing Map,” Springer-Verlag, 1995Google Scholar
  2. [2]
    S. W. Lee and H. H. Song, “LVQ Combined with Simulated Annealing for Optimal Design of Large-set Reference Models,” Neural Networks, Vol. 9, No. 2, pp. 329–336, 1996CrossRefGoogle Scholar
  3. [3]
    B. H. Juang and S. Katagiri, “Discriminative Learning for Minimum Error Classification,” IEEE Trans. on Signal Proc., Vol. 40, No. 12, pp. 3043–3054, 1992MATHCrossRefGoogle Scholar
  4. [4]
    A. Sato and K. Yamada, “Generalized Learning Vector Quantization,” In Advances in Neural Information Processing Systems 8, pp. 423–429, MIT Press, 1996Google Scholar

Copyright information

© Springer-Verlag London 1998

Authors and Affiliations

  • Atsushi Sato
    • 1
  • Keiji Yamada
    • 1
  1. 1.C&C Media Research LaboratoriesNEC CorporationKawasaki, KanagawaJapan

Personalised recommendations