ROBUST BAYESIAN-ANALYSIS UNDER GENERALIZED MOMENTS CONDITIONS

被引：19

作者：

BETRO, B

RUGGERI, F

MECZARSKI, M

机构：

[1] CNR,IST APPLICAZ MATEMAT & INFORMAT,I-20131 MILAN,ITALY

[2] WARSAW SCH ECON,INST ECONOMETR,PL-02554 WARSAW,POLAND

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 1994年 / 41卷 / 03期

关键词：

BAYESIAN ROBUSTNESS; POSTERIOR FUNCTIONALS; CLASSES OF PRIORS; GENERALIZED MOMENTS CONDITIONS; MARGINAL DISTRIBUTIONS;

D O I：

10.1016/0378-3758(94)90022-1

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

By applying the generalized moments theory we can construct the set of extreme points for a family of probability measures. Posterior functionals which are considered in Bayesian analysis attain their extremes at such points. This is the way to transform optimization with respect to measures into optimization with respect to points in a Euclidean space. The method is applied to known problems of Bayesian robustness as well as to a new class of priors, which is defined by conditions on the marginal distributions of data.

引用

页码：257 / 266

页数：10

共 8 条

[1] ROBUST BAYES AND EMPIRICAL BAYES ANALYSIS WITH EPSILON-CONTAMINATED PRIORS [J].