Bias reduction, the Jeffreys prior and GLIM

  • David Firth
Part of the Lecture Notes in Statistics book series (LNS, volume 78)

Summary

In full exponential-family models, such as generalized linear models with canonical link, use of the Jeffreys invariant prior as a penalty function removes the leading term from the asymptotic bias of maximum likelihood estimates. In Poisson log linear and binomial logistic linear models, the posterior mode can be calculated using the standard iterative weighted least-squares algorithm with an appropriate adjustment to the working vector at each cycle. The required adjustment is a simple function of case leverages. Implementation is particularly straightforward in GLIM4, and illustrates the power of the new ‘OWN algorithm’ facility.

Keywords

Covariance No_go 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Copas, J. B. (1988). Binary regression models for contaminated data. J. R. Statist. Soc. B 50, 225–265.MathSciNetGoogle Scholar
  2. Cordeiro, G. M. & McCullagh, P. (1991). Bias correction in generalized linear models. J. R. Statist. Soc. B 53, 629–643.MathSciNetMATHGoogle Scholar
  3. Firth, D. (1991). Bias reduction of maximum likelihood estimates. Preprint no. 209, Department of Mathematics, University of Southampton.Google Scholar
  4. Firth, D. (1992). Generalized linear models and Jeffreys priors: an iterative weighted least-squares approach. To appear in the proceedings of COMPSTAT 92. Physica-Verlag.Google Scholar
  5. Jeffreys, H. (1946). An invariant form for the prior probability in estimation problems. Proc. Roy. Soc. A 186, 453–461.MathSciNetMATHCrossRefGoogle Scholar
  6. McCullagh, P. & Nelder, J. A. (1989). Generalized Linear Models. 2nd Edition. Chapman and Hall, London.MATHGoogle Scholar
  7. Tierney, L. (1990). LISP-STAT: An Object-Oriented Environment for Statistical Computing and Dynamic Graphics. Wiley, New York.MATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag New York, Inc. 1992

Authors and Affiliations

  • David Firth
    • 1
  1. 1.Department of MathematicsUniversity of SouthamptonEngland

Personalised recommendations