Advertisement

Small Size Designs in Nonlinear Models Computed by Stochastic Optimization

  • J.-P. Gauchi
  • A. Pázman
Conference paper
Part of the Contributions to Statistics book series (CONTRIB.STAT.)

Summary

Optimality criteria, that are functions of the mean square error matrix, are expressed as integrals of the density of the parameter estimator. The optimum design is obtained by an accelerated stochastic optimization method. The estimator is modified to reflect prior knowledge about the parameters, and to take into account the boundary of the parameter space. Results of (1992) are extended and improved by that. Computer results are presented on examples.

Key words

A- and D-optimality distribution of estimators Mean Square Error 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. J. Baranyi and T.A. Roberts. A dynamic approach to predicting bacterial growth in food. Int. J. Food. Microbiol.23:277–2941994.CrossRefGoogle Scholar
  2. J. Baranyi and T.A. Roberts. Mathematics of predictive food microbiology. Int. J. Food. Microbiol.26:199–2181995.CrossRefGoogle Scholar
  3. J.R. Blum. Multidimensional stochastic approximation methods. Ann. Stat.25:737–7441954.zbMATHCrossRefGoogle Scholar
  4. M. Clyde and K. Chaloner. Constrained designs strategies for improving normal approximations in nonlinear regression problems. J. Stat. Plann. Inf.104(1):175–1962002.MathSciNetzbMATHCrossRefGoogle Scholar
  5. V. Fabian. Stochastic approximation of minima with improved asymptotic speed. Ann. Stat.38:191–2001967.Google Scholar
  6. J.R Gauchi. Contribution à la l’étude et au calcul de critères de plans d’expériences optimaux pour des modèles de régression non linéaire. PhD thesis, CNAM, Paris1999.Google Scholar
  7. J.P. Gauchi and A. Pázman. Distribution of the least squares estimator, stochastic, optimization and optimum designs in nonlinear models. Technical Report 2003-7, INRA, Biometrics Unit, Jouy-en-Josas, 2003.Google Scholar
  8. J.P. Gauchi and J.P. Vila. Réduction de variance par échantillonnage selon l’importance: application à l’optimisation stochastique d’un critère de plans d’expériences. In Actes des XXVIèmes Journées de Statistique ASU, Neuchatel, 1994, pages 317–320. ASU, 1994.Google Scholar
  9. A. Pázman. Probability distribution of the multivariate nonlinear least squares estimates. Kybernetika20:209–2301984.MathSciNetzbMATHGoogle Scholar
  10. A. Pázman. Higher dimensional nonlinear regression — a statistical use of the Riemannian curvature tensor. Statistics25:17–251993.MathSciNetzbMATHCrossRefGoogle Scholar
  11. A. Pázman. Nonlinear statistical models, volume 254 of Mathematics and its Applications. Kluwer Academic Publishers Group, Dordrecht1993.Google Scholar
  12. A. Pázman and L. Pronzato. Nonlinear experimental design based on the distribution of estimators. J. Stat Plann. Inf.33:385–4021992.zbMATHCrossRefGoogle Scholar
  13. G. Ch. Pflug. Stepsize rules, stopping times and their implementation in stochastic quasigradient algorithms. In Yu. Ermoliev and R.J.-B. Wets, editorsNumerical techniques for stochastic optimization, pages 353–372. Springer-Verlag, Heidelberg1988.CrossRefGoogle Scholar
  14. B.T. Polyak and Ya. Z. Tsypkin. Pseudogradient adaptation and training algorithms. Aut. Remote Control34(3):377–3971973.MathSciNetGoogle Scholar
  15. L. Pronzato. Synthèse d’expériences robustes pour modèles à paramètres incertains. PhD thesis, Université de Paris-Sud, Orsay1986.Google Scholar
  16. J.-P. Vila. Exact experimental designs via stochastic optimization for nonlinear regression models. In COMPSTAT 1990 (Dubrovnik, 1990), pages 291–296Heidelberg1990. Physica.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • J.-P. Gauchi
    • 1
  • A. Pázman
    • 2
  1. 1.Biometrics UnitINRA-Jouy en JosasJouy-en-Josas cedexFrance
  2. 2.Comenius UniversityBratislava

Personalised recommendations