Open Conference Systems, 50th Scientific meeting of the Italian Statistical Society

Font Size: 
Posterior distributions with non explicit objective priors
Erlis Ruli, Nicola Sartori, Laura Ventura

Last modified: 2018-08-30

Abstract


We introduce two methods useful to derive a posterior distribution for a parameter of interest, when only the first derivative of a log-prior is available. This is typically the situation when dealing with multidimensional parameters and objective priors. An example is illustrated using a predictive matching prior.



References


  1. Chernozbukov,V.,Hong,H.:AnMCMCapproachtoclassicalestimation,J.Econometrics115, 1234–1241 (2003)

  2. Corcuera, J.M., Giummole`, F.: A generalized Bayes rule for prediction. Scand. J. Statist. 26, 265–279 (1999)

  3. Datta, G.S., Mukerjee, R.: Probability Matching Priors: Higher-Order Asymptotics. Lecture Notes in Statistics, Springer (2004)

  4. Firth,D.:Biasreductionofmaximumlikelihoodestimates.Biometrika80,27–38(1993)

  5. Heinze,G.,Schemper,M:ASolutiontotheproblemofseparationinlogisticregression.Statist.

    Med. 21, 2409–2419 (2002)

  6. Komaki, F.: On asymptotic properties of predictive distributions. Biometrika 83, 299–313

    (1996)



Full Text: PDF