Advertisement

ANN Realizations of Local Approximations Schemes

  • A. A. Pervozvanskii
Conference paper

Abstract

Some schemes of the local approximation (LA) of functions defined on multidimensional space are described. For the simplest variants, namely, for the space filtration scheme and Nadraja-Watson scheme, the approximation precision is estimated. It is shown that ANN based on LA can solve the approximation problem with a good precision if the space dimension is moderate and some special elements are used.

Keywords

Space Dimension Local Approximation Special Element Multidimensional Space Perception Structure 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Barron, A.: I2SE Tr.,v.IT-39,930(1993)Google Scholar
  2. 2.
    Katkovnik, V.; Non-parametric identification and data smoothing: method of local approximation,M.: Naulca 1985Google Scholar
  3. 3.
    Bakhvalov, N.: Vestnik MGU, Matser., NO5,3,(I959)Google Scholar
  4. 4.
    Gurney,K.: Neural Networks,v. 5, 289 (1992)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag/Wien 1995

Authors and Affiliations

  • A. A. Pervozvanskii
    • 1
  1. 1.Dept. of Mechanics and Contr.ProcessesSPb State Technical University 29St.PetersburgRussia

Personalised recommendations