Bayesian Analysis

Nonparametric elicitation for heavy-tailed prior distributions

John Paul Gosling, Anthony O'Hagan, and Jeremy E. Oakley

Full-text: Open access

Abstract

In the context of statistical analysis, elicitation is the process of translating someone's beliefs about some uncertain quantities into a probability distribution. The person's judgements about the quantities are usually fitted to some member of a convenient parametric family. This approach does not allow for the possibility that any number of distributions could fit the same judgements.

In this paper, elicitation of an expert's beliefs is treated as any other inference problem: the facilitator of the elicitation exercise has prior beliefs about the form of the expert's density function, the facilitator elicits judgements about the density function, and the facilitator's beliefs about the expert's density function are updated in the light of these judgements. This paper investigates prior beliefs about an expert's density function and shows how many different types of judgement can be handled by this method.

This elicitation method begins with the belief that the expert's density will roughly have the shape of a $t$~density. This belief is then updated through a Gaussian process model using judgements from the expert. The method gives a framework for quantifying the facilitator's uncertainty about a density given judgements about the mean and percentiles of the expert's distribution. A property of Gaussian processes can be manipulated to include judgements about the derivatives of the density, which allows the facilitator to incorporate mode judgements and judgements on the sign of the density at any given point. The benefit of including the second type of judgement is that substantial computational time can be saved.

Article information

Source
Bayesian Anal. Volume 2, Number 4 (2007), 693-718.

Dates
First available in Project Euclid: 22 June 2012

Permanent link to this document
https://projecteuclid.org/euclid.ba/1340370711

Digital Object Identifier
doi:10.1214/07-BA228

Mathematical Reviews number (MathSciNet)
MR2361971

Zentralblatt MATH identifier
1331.62217

Keywords
Expert elicitation Gaussian process heavy-tailed distribution nonparametric density estimation

Citation

Gosling, John Paul; Oakley, Jeremy E.; O'Hagan, Anthony. Nonparametric elicitation for heavy-tailed prior distributions. Bayesian Anal. 2 (2007), no. 4, 693--718. doi:10.1214/07-BA228. https://projecteuclid.org/euclid.ba/1340370711


Export citation

References

  • Ang, M. and Buttery, N. (1997). "An approach to the application of subjective probabilities in level 2 PSAs." Reliability Engineering and System Safety, 58: 145–156.
  • Belyaev, L. (1959). "Analytic random processes." Theor. Probability Appl., 4: 402–409.
  • Bernardo, J. and Smith, A. (1994). Bayesian Theory. Chichester: Wiley.
  • Blight, B. and Ott, L. (1975). "A Bayesian approach to model inadequacy for polynomial regression." Biometrika, 62: 79–88.
  • Cairns, J. and Shackley, P. (1999). "What price information? Modelling threshold probabilities of fetal loss." Social Science and Medicine, 49: 823–830.
  • Chaloner, K. and Rhame, F. (2001). "Quantifying and documenting prior beliefs in clinical trials." Statistics in Medicine, 20: 581–600.
  • Cooke, R. and Slijkhuis, K. (2003). "Expert Judgment in the Uncertainty Analysis of Dike Ring Failure Frequency." In Case Studies in Reliability and Maintenance (eds. W.R. Blischke and D.N. Prabhakar Murthy), 331–352. Chichester: Wiley.
  • Coolen, F., Mertens, P., and Newby, M. (1992). "A Bayes-competing risk model for the use of expert judgement in reliability estimation." Reliability Engineering and System Safety, 35: 23–30.
  • Daneshkhah, A., Oakley, J., and O'Hagan, A. (2006). "Nonparametric prior elicitation with imprecisely assessed probabilities." Technical report, BEEP working report, Department of Probability and Statistics, University of Sheffield. (Available from http://www.shef.ac.uk/beep/publications.html).
  • Dominitz, J. (1998). "Earning expectations, revisions, and realisations." The Review of Economics and Statistics, 80: 374–388.
  • Frijters, M., Cooke, R., Slijkhuis, K., and Noortwijk, J. V. (1999). "Expertmening Onzekerheidsanalyse." Technical report, Ministerie van verkeer en Waterstaat, Directoraat-Generaal Rijkswaterstaat, Bouwdienst Rijkswaterstaat ONIN-1-99006.
  • Garthwaite, P., Kadane, J., and O'Hagan, A. (2005). "Statistical methods for eliciting probability distributions." J. Amer. Statist. Ass., 100: 680–701.
  • Grisley, W. and Kellogg, E. (1983). "Farmers' subjective probabilities in Northern Thailand: an elicitation analysis." American Journal of Agricultural Economics, 77: 74–82.
  • Gustafson, D., Sainfort, F., Eichler, M., Adams, L., Bisognamo, M., and Steudel, H. (2003). "Developing and testing a model to predict outcomes of organizational change." Health Services Research, 38: 751–776.
  • Hogarth, R. (1975). "Cognitive processes and the assessment of subjective probability distributions." J. Amer. Statist. Ass., 70: 271–294.
  • Horrace, W. (2005). "Some results on the multivariate truncated normal distribution." Journal of Multivariate Analysis, 94: 209–221.
  • Kadane, J. and Wolfson, L. (1998). "Experiences in elicitation." The Statistician, 47: 3–19.
  • Kahneman, D. and Tversky, A. (1973). "On the psychology of prediction." Psychological review, 80: 237–251.
  • Kimeldorf, G. and Wahba, G. (1970). "A correspondence between Bayesian estimation on stochastic processes and smoothing by splines." Ann. Math. Statist., 41: 495–502.
  • Must, A., Phillips, S., Stunkard, A., and Naumova, E. (2002). "Expert opinion on body mass index percentiles for figure drawings at menarche." International Journal of Obesity, 26: 876–879.
  • Oakley, J. and O'Hagan, A. (2002). "Bayesian inference for the uncertainty distribution of computer model outputs." Biometrika, 89: 769–784.
  • –- (2007). "Uncertainty in prior elicitations." Technical report, 521/02, Department of Probability and Statistics, University of Sheffield. To appear in Biometrika.
  • O'Hagan, A. (1978). "Curve fitting and optimal design for prediction (with discussion)." J. R. Statist. Soc. Ser. B, 40: 1–42.
  • –- (1991). "Bayes-Hermite quadrature." J. Statist. Planning and Inference, 91: 245–260.
  • –- (1992). "Some Bayesian numerical analysis." In Bayesian Statistics 4 (eds. Bernado, J.M. et al.), 345–363. Oxford: Oxford University Press.
  • O'Hagan, A., Buck, C., Daneshkhah, A., Eiser, J., Garthwaite, P., Jenkinson, D., Oakley, J., and Rakow, T. (2006). Uncertain judgements: eliciting expert probabilities. Chichester: Wiley.
  • O'Hagan, A. and Leonard, T. (1976). "Bayes estimation subject to uncertainty about parameter constraints." Biometrika, 63: 201–203.
  • Ramachandran, G., Banerjee, S., and Vincent, J. (2003). "Expert judgement and occupational hygiene: application to aerosol speciation in the nickel primary production industry." Annals of Occupational Hygiene, 47: 461–476.
  • Rasmussen, C. (1996). "Evaluation of Gaussian processes and other methods for non-linear regression." Ph.D. thesis, Graduate Department of Computer Science, University of Toronto.
  • Sexsmith, R. (1999). "Probability-based safety analysis –- value and drawbacks." Structural Safety, 21: 303–310.
  • Smith, J. and Mandac, A. (1995). "Subjective versus objective yield distributions as measures of production risk." American Journal of Agricultural Economics, 77: 152–161.
  • Wernisch, L. (2004). "Nonlinear regression with Gaussian processes." Seminar Slides, Birbeck College, London. (Available from http://people.cryst.bbk.ac.uk/~wernisch/).