On Divergence Measures Leading to Jeffreys and Other Reference Priors

被引:9
作者
Liu, Ruitao [1 ,2 ]
Chakrabarti, Arijit
Samanta, Tapas
Ghosh, Jayanta. K. [3 ]
Ghosh, Malay [4 ]
机构
[1] Univ Iowa, Dept Stat & Actuarial Sci, Iowa City, IA 52242 USA
[2] ACT Inc, Iowa City, IA USA
[3] Purdue Univ, Dept Stat, W Lafayette, IN 47907 USA
[4] Univ Florida, Dept Stat, Gainesville, FL 32611 USA
来源
BAYESIAN ANALYSIS | 2014年 / 9卷 / 02期
关键词
alpha-divergences; Jeffreys prior; Reference prior; EXPONENTIAL-FAMILIES; INFORMATION; DISTRIBUTIONS; ASYMPTOTICS; PARAMETER;
D O I
10.1214/14-BA862
中图分类号
O1 [数学];
学科分类号
0701 ; 070101 ;
摘要
The paper presents new measures of divergence between prior and posterior which are maximized by the Jeffreys prior. We provide two methods for proving this, one of which provides an easy to verify sufficient condition. We use such divergences to measure information in a prior and also obtain new objective priors outside the class of Bernardo's reference priors.
引用
收藏
页码:331 / 369
页数:39
相关论文
共 45 条
[1]  
ALI SM, 1966, J ROY STAT SOC B, V28, P131
[3]  
[Anonymous], 2003, BAYESIAN NONPARAMETR
[4]  
[Anonymous], 1983, MULTIPLE INTEGRALS C
[5]  
[Anonymous], 2012, Differential-geometrical methods in statistics
[6]  
[Anonymous], 1994, Higher Order Asymptotics
[7]  
[Anonymous], 1992, BAYESIAN STAT
[8]  
BARTLETT MS, 1953, BIOMETRIKA, V40, P306
[9]  
Berger J. O., 1992, Bayesian Statistics, V4, P35
[10]   THE FORMAL DEFINITION OF REFERENCE PRIORS [J].
Berger, James O. ;
Bernardo, Jose M. ;
Sun, Dongchu .
ANNALS OF STATISTICS, 2009, 37 (02) :905-938