Bayesian inference and the parametric bootstrap Efron 2013外文电子书籍.pdfVIP

The Annals of Applied Statistics 2012, Vol. 6, No. 4, 1971–1997 DOI: 10.1214/12-AOAS571 c Institute of Mathematical Statistics, 2012 BAYESIAN INFERENCE AND THE PARAMETRIC BOOTSTRAP By Bradley Efron1 3 1 Stanford University 0 2 The parametric bootstrap can be used for the efficient computa- n tion of Bayes posterior distributions. Importance sampling formulas a take on an easy form relating to the deviance in exponential fami- J lies and are particularly simple starting from Jeffreys invariant prior. 4 Because of the i.i.d. nature of bootstrap sampling, familiar formulas 1 describe the computational accuracy of the Bayes estimates. Besides computational methods, the theory provides a connection between ] Bayesian and frequentist analysis. Efficient algorithms for the fre- P quentist accuracy of Bayesian inferences are developed and demon- A strated in a model selection example. . t a t 1. Introduction. This article concerns the use of the parametric boot- s [ strap to carry out Bayesian inference calculations. Two main points are 1 made: that in the comparatively limited set of cases where bootstrap meth- v ods apply,