Bayesian methods in extreme value modelling: A review and new developments
Bayesian methods in extreme value modelling: A review and new developments
复制标题
DOI：
10.2307/1403426
复制
发表时间：
1996-04-01
影响因子：
2
通讯作者：
Powell, EA
中科院分区：
文献类型：
--
作者：
Coles, SG;Powell, EA
极值问题的特点是数据稀缺，并且需要在数据最稀疏的地方建模，这在考虑贝叶斯推理方法时提出了一个困境：附加的先验信息的价值可能是相当大的，但是将这种先验知识公式化用于极端行为的可行性是值得怀疑的，在本文中，我们回顾了贝叶斯和极值分析的文献，并使用贝叶斯计算工具的最新进展来评估贝叶斯极值分析在三种不同情况下的效用：第一种是专家可以提供先验信息;第二种是最大似然法失败;第三种是相关变量的空间信息用于制定经验先验。
Extreme value problems are characterized by a scarcity of data and the requirement of modelling where the data are most sparse, This presents a dilemma when considering a Bayesian approach to inference: the value of additional prior information is likely to be substantial, but the plausibility of formulating such prior knowledge for extremal behaviour is questionable, In this paper we review the literature linking the themes of Bayesian and extreme value analysis, and use recent advances in Bayesian computational tools to assess the utility of a Bayesian extreme value analysis in three different situations: one where an expert is available to supply prior information; the second where maximum likelihood fails; and the third where spatial information on related variables is used to formulate an empirical prior.