Noninformative priors for the quantile of lognormal distribution
Abstract
We development noninformative priors for the quantiles of the log-normal distribution in this article. The priors are obtained by matching the coverage probabilities of one-sided Bayesian confidence intervals with the corresponding frequentist coverage probabilities. In addition, the reference priors are derived. The reference prior matches the second-order matching prior, but Jeffreys’ prior is not. We show that the second-order matching prior is the only prior that satisfies various matching criteria. Sufficient conditions for the adequacy of the posterior distribution under general priors including reference priors are presented. We perform a simulation study and provide a real example.
Fields of Study & Subjects
Statistical Distribution Estimation and Applications
Statistical Methods and Bayesian Inference
Probability and Statistical Research
Publication details
- Journal / source
- Journal of the Korean Data and Information Science Society
- Volume
- 37
- Issue
- 1
- Pages
- 139–149
- Publication date
- 31 January 2026
- Language
- EN