Bayesian Bootstrap Confidence Intervals for Population Quantiles and Empirical Comparisons

MUSIO, MONICA;TARGHETTA M. L.

2006-01-01

---

Abstract

Rubin's method (Rubin 1981) is applied to construct Bayesian bootstrap confidence intervals for the quantiles of an unknown distribution function. Comparisons with bootstrap percentile confidence intervals are carried out by Monte Carlo experiments relative to an exponential probability distribution with an unknown shift parameter.