BAYESIAN-INFERENCE OF BINARY REGRESSION-MODELS WITH PARAMETRIC LINK

被引:16
|
作者
CZADO, C [1 ]
机构
[1] YORK UNIV,DEPT MATH & STAT,N YORK M3J 1P3,ON,CANADA
基金
加拿大自然科学与工程研究理事会;
关键词
BAYESIAN INFERENCE; BINARY RESPONSE MODELS; LINK TRANSFORMATIONS; PROBIT REGRESSION; MARKOV CHAIN MONTE-CARLO SAMPLING METHODS; GIBBS SAMPLER; METROPOLIS ALGORITHM; REJECTION SAMPLING;
D O I
10.1016/0378-3758(94)90158-9
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
A computationally tractable Bayesian inference for binary regression models with parametric link is derived utilizing a Markov chain Monte Carlo algorithm to simulate from the joint posterior distribution of the regression and the fink parameter. In particular, Bayesian posterior calculations are faciltated for a generalized probit regression model involving tail modification by developing a hybrid sampling algorithm with Gibbs and Metropolis/rejection sampling steps. It also involves sufficiency and other reduction arguments to optimize the algorithm. The proposed algorithm is applied to two examples producing marginal and joint posterior density estimates for link and regression parameters. Bayesian point and interval estimates for these parameters as well as for the corresponding success probabilities are also investigated.
引用
收藏
页码:121 / 140
页数:20
相关论文
共 50 条