Advertisement

Identification and prediction of non-linear models with recurrent neural network

  • Adam Olivier
  • Zarader Jean-Luc
  • Milgram Maurice
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 686)

Abstract

Using a neural network to identify models and predict signals allows to go beyond the linear domain. In this paper, we show the advantage of using neural network for these signal processing applications. Thus, the function charactering the cell (sigmoïd or others) allows the study of non-linear models. Using feedback links specific to a recurrent network, the time is taken into account. Two different goals are assigned to the two phases in using this type of network: 1) the neural network training method uses a gradient backward propagation method. During the learning phase, the weights of the network are modified to identify the parameters of the given model. 2) during the test phase, the network predicts the output for each time step. Results are presented in the case of a Non-Linear AutoRegressive filters and they confirm the good responses of neural networks both for identification of parameters and for prediction of output for these non-linear models.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    D. Rumelhart, G.E. Hintonand R.J. Williams, “Learning Internal Representations by Error Propagation”, Parallel Distributed Processing: Explorations in the Microstructure of Cognition. MIT Press.Google Scholar
  2. [2]
    E. Levin, “A Recurrent Neural Network: limitations and training”, Neural Networks, vol. 3, pp. 641–650, 1990.Google Scholar
  3. [3]
    J.L. Elman, “Finding Structure in Time”, Cognitive Science, n∘14, pp. 179–211, 1990.Google Scholar
  4. [4]
    Y. Le Cunn, “Modèles Connexionnistes de l'Apprentissage”, Thése de l'Université Paris 6, 1987Google Scholar
  5. [5]
    Ronald J. Williams and David Zisper, “Experimental Analysis of Real-Time Recurrent Learning Algorithm”, Connection Science, Vol.1, n∘1, 1989.Google Scholar
  6. [6]
    F.J. Pineda, “Generalization of back propagation to recurrent networks”, Phys. Rev. Lett., vol.59, n∘19, pp.2229–2232, Nov.87.Google Scholar
  7. [7]
    B. Pearlmutter, “Dynamic Recurrent Neural Networks”, School of Computer Science, Carnegie Melln University, Pittsburgh, PA 15 213, Déc.90.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1993

Authors and Affiliations

  • Adam Olivier
    • 1
  • Zarader Jean-Luc
    • 1
  • Milgram Maurice
    • 1
  1. 1.Laboratoire de Robotique de ParisURA 1305Paris Cedex 05France

Personalised recommendations