Advertisement

On Improved Performance Index Function with Enhanced Generalization Ability and Simulation Research

  • Dongcai Qu
  • Rijie Yang
  • Yulin Mi
Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 15)

Abstract

It was one of most main performance that identification model of Artificial Neural Network (ANN) was Generalization Ability. It also was one of key question researched by domestic and foreign concerned experts in the recent years. Generalization ability of ANN’s identification model had concerned with many factors, and appropriate designed performance index function was an important influence factor. After common performance index function was analyzed based on the mean error function smallest principle, a kind of improved performance index function was obtained through joined the power values to the time delay information. The massive simulation researches show that improved performance index function is effective to enhance generalization ability of ANN’s models.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Wei, H.K.: Theory and Method of Framework Design of the Neural Networks. National Defense Industry Press, Beijing (2001)Google Scholar
  2. 2.
    Yuan, P.F.: Complexity of Capability and Learning and Computation of the Neural Networks. University of Tsinghua, Beijing (2001)Google Scholar
  3. 3.
    Yuan, P.F., Zhang, N.Y.: Neural networks and analog evolvement computation. University of Tsinghua, Beijing (2001)Google Scholar
  4. 4.
    Zhang, N.Y., Yuan, P.F.: Neural Networks and Fuzzy Control. University of Tsinghua, Beijing (1998)Google Scholar
  5. 5.
    Larsen, J., Hansen, L.K.: Generalization Performance of Regularized Neural Network Models. In: Proc. of the IEEE Workshop on Neural networks for Signal Proc. IV, Piscataway, New Jersey, pp. 42–51 (1994)Google Scholar
  6. 6.
    Mass, W.: Neural Nets with superlinear VC-dimension. Neural Computation, 877–884 (1994)Google Scholar
  7. 7.
    Girosi, F., Jones, M., Poggio, T.: Regularization Theory and Neural Network Architecture. Neural Computation, 219–269 (1995)Google Scholar
  8. 8.
    Williams, P.M.: Bayesian Regularization and Pruning using a Laplace Prior. Neural Computation, 117–143 (1995)Google Scholar
  9. 9.
    Sjoberg, J., Ljung, L.: Overtraining, Regularization, and Searching for Minimum in Neural Networks. In: Preprint IFAC Symp. on Adaptive Systems in Control and Signal Processing, Grenoble, France, pp. 669–674 (1992)Google Scholar
  10. 10.
    Mackay, D.: A Practical Bayesian framework for Backpropagation Networks. Neural Computation, 448-472 (1992)Google Scholar
  11. 11.
    Yang, O., Lin, Q.: The Discussion on Improving Generalization Ability of a Feedforward Neural Network. Journal of Nanpin Teachers College, 60–63 (2007)Google Scholar
  12. 12.
    Sjoberg, J., Ljung, L.: Overtraining, Regularization, and Searching for a Minimum, With Application to Neural Networks. International Journal of Control, 1391–1407 (1995)Google Scholar
  13. 13.
    Bishop, C.M.: Training with Noise is Equivalent to Tikhonov Regularization. Neural Computation, 108–116 (1995)Google Scholar
  14. 14.
    An, G.: The Effect of Adding Noise during Backpropagation Training on a Generalization Performance. Neural Computation, 643–671 (1996)Google Scholar
  15. 15.
    Krogh, A., Hertz, J.: A Simple Weight Decay Can Improve Generalization. NIPS 4, 950–957 (1992)Google Scholar
  16. 16.
    Norgaard, M.: Neural Network Based System Identification Toolbox, Ver. 2, Tech. Report 00-E-891, Department of Automation, Technical University of Denmark (2000)Google Scholar
  17. 17.
    Norgaard, M., Ravn, O., Poulsen, N.K.: Neural Networks for Modeling and Control of Dynamic Systems. Springer, London (2000)Google Scholar
  18. 18.
    Ljung, L.: System Identification –Theory for the User. Prentice-Hall, Englewood Cliffs (1987)zbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Dongcai Qu
    • 1
  • Rijie Yang
    • 2
  • Yulin Mi
    • 3
  1. 1.Department of Control Engineering of NavalAeronautical and Astronautical UniversityYantaiP.R. China
  2. 2.Department of Electronic and Information Engineering of NavalAeronautical and Astronautical UniversityYantaiP.R. China
  3. 3.Department of Training of NavalAeronautical and Astronautical UniversityYantaiP.R. China

Personalised recommendations