A Novel Recursive Algorithm for Training RBF Networks

Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 288)

Abstract

A recursive learning algorithm is presented for basis selection of radial basis function (RBF) neural network. It is based on an adaptive kernel width algorithm, which can select basis functions recursively in the nonorthogonal space and assign an appropriate number of hidden units of RBF network. This also makes the model structure independent of the selected term sequence and assures an optimal RBF network even if the RBF original basis is nonorthogonal. Its effectiveness is demonstrated by the simulated results.

Keywords

Radial basis function Bayesian information criterion Recursive 

References

  1. 1.
    Huang GB, Saratchandran P, Sundararajan N (2005) A generalized growing and pruning RBF (GGAP-RBF) neural network for function approximation. IEEE Trans Neural Network 16(1):57–67CrossRefGoogle Scholar
  2. 2.
    Chen S, Cowan CFN, Grant PM (1991) Orthogonal least squares learning algorithm for radial basis function networks. IEEE Trans Neural Network 2(2):302–309CrossRefGoogle Scholar
  3. 3.
    Sherstinsky A, Picard RW (1996) On the efficiency of the orthogonal least squares training method for radial basis function networks. IEEE Trans Neural Network 7(1):195–200CrossRefGoogle Scholar
  4. 4.
    Li K, Peng JX, Bai EW (2009) Two-stage mixed discrete-continuous identification of radial basis function (RBF) neural model for nonlinear systems. IEEE Trans Circuits Syst 56(3):630–643MathSciNetCrossRefGoogle Scholar
  5. 5.
    Deng J, Li K, Irwin GW (2010) A two-stage algorithm for automatic construction of RBF neural model. In: 15th IEEE mediterranean electrotechnical conference, vol 5795-9. pp 166–171Google Scholar
  6. 6.
    Schwarz G (1978) Estimating the dimension of a model. Ann Stat 6(2):461–464CrossRefMATHGoogle Scholar
  7. 7.
    Li K, Peng J, Irwin GW (2005) A fast nonlinear model identification method. IEEE Trans Autom Control 50(8):1211–1216MathSciNetCrossRefGoogle Scholar
  8. 8.
    Peng Z, Dehua L, Hong W, Feng C (2011) The automatic model selection and variable width for RBF neural networks. Neurocomputing 74:3628–3637Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  1. 1.State Key Laboratory of Optical Communication Technologies and NetworksWuhan Research Institute of Posts and TelecommunicationsWuhanChina

Personalised recommendations