Statistics and Computing

, Volume 19, Issue 2, pp 189–201 | Cite as

Bayesian estimation of quantile distributions

Article

Abstract

Use of Bayesian modelling and analysis has become commonplace in many disciplines (finance, genetics and image analysis, for example). Many complex data sets are collected which do not readily admit standard distributions, and often comprise skew and kurtotic data. Such data is well-modelled by the very flexibly-shaped distributions of the quantile distribution family, whose members are defined by the inverse of their cumulative distribution functions and rarely have analytical likelihood functions defined. Without explicit likelihood functions, Bayesian methodologies such as Gibbs sampling cannot be applied to parameter estimation for this valuable class of distributions without resorting to numerical inversion. Approximate Bayesian computation provides an alternative approach requiring only a sampling scheme for the distribution of interest, enabling easier use of quantile distributions under the Bayesian framework. Parameter estimates for simulated and experimental data are presented.

Keywords

Approximate Bayesian computation Posterior distribution Quantile distribution Response time data 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Alston, C., Mengersen, K., Thompson, J., Littlefield, P., Perry, D., Ball, A.: Statistical analysis of sheep CAT scan images using a Bayesian mixture model. Austr. J. Agric. Res. 55(1), 57–68 (2004) CrossRefGoogle Scholar
  2. Beaumont, M., Zhang, W., Balding, D.: Approximate Bayesian computation in population genetics. Genetics 162, 2025–2035 (2002) Google Scholar
  3. Excoffier, L., Estoup, A., Cornuet, J.-M.: Bayesian analysis of an admixture model with mutations and arbitrarily linked markers. Genetics 169, 1727–1738 (2005) CrossRefGoogle Scholar
  4. Freimer, M., Mudholkar, G., Kollia, G., Lin, C.: A study of the generalized Tukey-lambda family. Commun. Stat. Theory Methods 17(10), 3547–3567 (1988) MATHCrossRefMathSciNetGoogle Scholar
  5. Gelman, A., Carlin, J., Stern, H., Rubin, D.: Bayesian Data Analysis. Chapman & Hall/CRC, Boca Raton (2004) MATHGoogle Scholar
  6. Haynes, M.: Flexible distributions and statistical models in ranking and selection procedures, with applications. Ph.D. Thesis, Queensland University of Technology (1998) Google Scholar
  7. Haynes, M., Mengersen, K.: Bayesian estimation of g-and-k distributions using MCMC. Comput. Stat. 20(1), 7–30 (2005) CrossRefMathSciNetGoogle Scholar
  8. Haynes, M., MacGillivray, H., Mengersen, K.: Robustness of ranking and selection rules using generalised g-and-k distributions. J. Stat. Plan. Inference 65(1), 45–66 (1997) MATHCrossRefMathSciNetGoogle Scholar
  9. Heathcote, A., Popiel, S., Mewhort, D.: Analysis of response time distributions: An example using the Stroop task. Psychol. Bull. 109(2), 340–347 (1991) CrossRefGoogle Scholar
  10. Hoeting, J., Madigan, D., Raftery, A., Volinsky, C.: Bayesian model averaging: A tutorial. Stat. Sci. 14(4), 382–401 (1999). A version with typesetting errors corrected is available at http://www.stat.colostate.edu/~jah/papers/statsci.pdf MATHCrossRefMathSciNetGoogle Scholar
  11. Kass, R., Raftery, A.: Bayes factors. J. Am. Stat. Assoc. 90, 773–795 (1995) MATHCrossRefGoogle Scholar
  12. King, R., MacGillivray, H.: A starship fitting method for the generalized lambda distribution. Aust. N. Z. J. Stat. 41(3), 353–374 (1999) MATHCrossRefMathSciNetGoogle Scholar
  13. Marin, J.-M., Robert, C.: Bayesian Core. Springer, New York (2007) MATHGoogle Scholar
  14. Marjoram, P., Molitor, J., Plagnol, V., Tavaré, S.: Markov chain Monte Carlo without likelihoods. Proc. Nat. Acad. Sci. USA 100(26), 15324–15328 (2003) CrossRefGoogle Scholar
  15. Pritchard, J., Seielstad, M., Perez-Lezaun, A., Feldman, M.: Population growth of human Y chromosomes: a study of Y chromosome microsatellites. Mol. Biol. Evol. 16(12), 1791–1798 (1999) Google Scholar
  16. Ramberg, J., Schmeiser, B.: An approximate method for generating asymmetric random variables. Commun. ACM 17(2), 78–82 (1974) MATHCrossRefMathSciNetGoogle Scholar
  17. Rayner, G., MacGillivray, H.: Numerical maximum likelihood estimation for the g-and-k and generalized g-and-h distributions. Stat. Comput. 12(1), 55–75 (2002a) CrossRefMathSciNetGoogle Scholar
  18. Rayner, G., MacGillivray, H.: Weighted quantile-based estimation for a class of transformation distributions. Comput. Stat. Data Anal. 39(4), 401–433 (2002b) MATHCrossRefMathSciNetGoogle Scholar
  19. Su, S.: Numerical maximum log likelihood estimation for generalized lambda distributions. Comput. Stat. Data Anal. 51(8), 3983–3998 (2007) CrossRefGoogle Scholar
  20. Tavaré, S., Balding, D., Griffiths, R., Donnelly, P.: Inferring coalescence times from DNA sequence data. Genetics 145(2), 505–518 (1997) Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2008

Authors and Affiliations

  • D. Allingham
    • 1
  • R. A. R. King
    • 2
  • K. L. Mengersen
    • 3
  1. 1.Centre for Complex Dynamic Systems and Control, School of Mathematical and Physical SciencesThe University of NewcastleCallaghanAustralia
  2. 2.School of Mathematical and Physical SciencesThe University of NewcastleCallaghanAustralia
  3. 3.School of Mathematical SciencesQueensland University of TechnologyBrisbaneAustralia

Personalised recommendations