BibTeX
@MISC{In_usefulnessof,
author = {Carlo Model In and M. Ribatet},
title = {Usefulness of the Reversible Jump Markov Chain Monte},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Regional flood frequency analysis is a convenient way to reduce estimation uncertainty 11 when few data are available at the gauging site. In this work, a model that allows a non 12 null probability to a regional fixed shape parameter is presented. This methodology is inte- 13 grated within a Bayesian framework and uses reversible jump techniques. The performance 14 on stochastic data of this new estimator is compared to two other models: a conventional 15 Bayesian analysis and the index flood approach. Results show that the proposed estimator is 16 absolutely suited to regional estimation when only a few data is available at the target site. 17 Moreover, Bayesian models appear to be more robust on error estimation on the target site 18 index flood estimation than the index flood estimator. Some suggestions about configurations 19 of the pooling groups are also presented to increase the performance of each estimator. 20 Keywords: Regional Frequency Analysis, Extreme Value Theory, Generalized Pareto Dis- 21 tribution, Reversible Jumps, Markov Chain Monte Carlo. 22 1
Keyphrases
reversible jump markov chain monte target site bayesian analysis reversible jump technique generalized pareto dis convenient way estimation uncertainty pooling group markov chain monte carlo reversible jump stochastic data regional frequency analysis error estimation regional fixed shape parameter regional estimation regional flood frequency analysis index flood estimation new estimator extreme value theory bayesian model null probability bayesian framework
