Nonlinear System Modeling Using Wavelet Networks

  • Seda Postalcioglu
  • Yasar Becerikli
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3497)

Abstract

In this paper we examine modeling of a nonlinear system using wavelet networks. Wavelet networks are similar to neural networks for the structure and the learning approach. But training algorithms for wavelet networks require a smaller number of iterations when compared with neural networks. Also interpretation of the model with neural networks is so hard. Gaussian based mother wavelet function is used as an activation function. Wavelet networks have these parameters; dilation, translation, and weights. Wavelets are rapidly vanishing functions. For this reason heuristic procedure has been used. Selecting initial values of weights are made randomly. Then parameters are optimized during learning. To update parameters, gradient method has been applied by using momentum. Quadratic cost function is used for error minimization. Two test data have been used for the simulations. One of them is a static function and the other one is a second order nonlinear function.

Keywords

Mother Wavelet Dynamic System Modeling Wavelet Frame Wavelet Network Quadratic Cost Function 
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.
    Zhang, Q., Benveniste, A.: Wavelet Networks. IEEE Transaction on Neural Networks 3, 889–898 (1992)CrossRefGoogle Scholar
  2. 2.
    Galvao Roberto, K.H., Becerra, V.M.: Linear-Wavelet Models for System Identification. In: IFAC15th Triennial World Congress, Barcelona, Spain (2002)Google Scholar
  3. 3.
    Zhang, Q.: Using Wavelet Network in Non parametric Estimation. IEEE trans. Neural Networks 8, 227–236 (1997)CrossRefGoogle Scholar
  4. 4.
    Thuillard, M.: Review of Wavelet Networks, Wavenets, Fuzzy Wavenets and Their Applications. In: ESIT 2000, Aachen, Germany (2000)Google Scholar
  5. 5.
    Polycarpou, M., Mears, M., Weaver, S.: Adaptive Wavelet Control of Nonlinear Systems. In: Proceedings of the1997 IEEE Conference on Decision and Control, pp. 3890–3895 (1997)Google Scholar
  6. 6.
    Pati, Y.C., Krishnaprasad, P.S.: Analysis and Synthesis of Feedforward Neural Networks Using Discrete Affine Wavelet Transformations. IEEE Trans. On Neural Networks 4, 73–85 (1993)CrossRefGoogle Scholar
  7. 7.
    Oussar, Y., Rivals, I., Personnaz, L., Dreyfus, G.: Training Wavelet Networks for Nonlinear Dynamic Input Output Modeling. Elsevier neurocomputing 20, 173–188 (1998)CrossRefMATHGoogle Scholar
  8. 8.
    Oussar, Y., Dreyfus, G.: Initialization by Selection for Wavelet Network Training. Neurocomputing 34, 131–143 (2000)CrossRefMATHGoogle Scholar
  9. 9.
    Reza, A.M.: Wavelet Characteristics. White Paper, Spire Lab. UWM (1999)Google Scholar
  10. 10.
    Polikar R.: The wavelet tutorial (2001), http://engineering.rowan.edu/~polikar/WAVELETS
  11. 11.
    Becerikli, Y., Oysal, Y., Konar, A.F.: On a Dynamic Wavelet Network and Its Modeling Application. In: Kaynak, O., Alpaydın, E., Oja, E., Xu, L. (eds.) ICANN 2003 and ICONIP 2003. LNCS, vol. 2714, pp. 710–718. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  12. 12.
    Becerikli, Y.: On Three Intelligent Systems: Dynamic Neural, Fuzzy and Wavelet Networks For Training Trajectory. Neural Computing & Applications (NC&A) 13, 339–351 (2004)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2005

Authors and Affiliations

  • Seda Postalcioglu
    • 1
  • Yasar Becerikli
    • 2
  1. 1.Technical Education Faculty, Electronic-Comp. Educ. DepartmentKocaeli UniversityIzmitTurkey
  2. 2.Computer Engineering DepartmentKocaeli UniversityIzmitTurkey

Personalised recommendations