Application of Mixture Priors to Assessment of Performance

  • Corwin L. Atwood
  • Robert W. Youngblood
Conference paper


A“mixture prior” offers significant advantages in performance assessment, compared to diffuse priors previously considered for this application. In particular, a discrete mixture of natural conjugate priors gives a simple and relatively realistic way to model past performance while allowing the possibility of detecting changes in performance. For a range of possible data in a specific example, the behaviour of the resulting posterior is compared to that of posteriors resulting from more traditional priors.


Maximum Likelihood Estimate Posterior Distribution Prior Distribution Past Performance Degraded State 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Box GEP and Tiao GC. Bayesian inference in statistical analysis. Addison-Wesley, Reading, MA, 1973Google Scholar
  2. 2.
    Atwood CL. Constrained noninformative priors in risk assessment. Rel. Eng. and Sys. Safety 1996; 53: 37–46CrossRefGoogle Scholar
  3. 3.
    Hamzehee HG et al. Risk-based performance indicators: results of phase 1 development. NUREG-1753, 2002Google Scholar
  4. 4.
    Eide SA. Historical perspective on failure rates for US commercial reactor components. Rel. Eng. and Sys. Safety 2003; 80: 123–132CrossRefGoogle Scholar
  5. 5.
    Atwood CL. Mixture priors for estimating risk. STATWOOD-2003/1, 2003Google Scholar
  6. 6.
    Dalai SR, Hall WJ. Approximating priors by mixtures of natural conjugate priors. J.Roy. Stat. Soc., Ser. B 1983; 45: 278–286Google Scholar
  7. 7.
    Diaconis P, Ylvisaker D. Quantifying prior opinion. In: Bernardo JM et al. (eds) Bayesian Statistics 2: Proceedings of the Second Valencia International Meeting. North-Holland, Amsterdam, 1985, pp 133–156Google Scholar

Copyright information

© Springer-Verlag London 2004

Authors and Affiliations

  • Corwin L. Atwood
    • 1
  • Robert W. Youngblood
    • 2
  1. 1.Statwood ConsultingSilver SpringUSA
  2. 2.ISL, Inc.RockvilleUSA

Personalised recommendations