Abstract
In 1981 Rubin introduced the Bayesian bootstrap and argued that it was the natural Bayesian analogue to the usual bootstrap. We show here that when estimating a population quantile in a nonparametric problem it yields estimators that are often preferred to the natural naive estimators based on the order statistic.
Original language | English (US) |
---|---|
Pages (from-to) | 103-109 |
Number of pages | 7 |
Journal | Statistics and Probability Letters |
Volume | 16 |
Issue number | 2 |
DOIs | |
State | Published - 1993 |
Bibliographical note
Funding Information:in part by NSF grant
Keywords
- Bayesian bootstrap
- noninformative Bayes
- nonparametric estimation
- order statistic
- quantile