Non-iterative Sampling Method for Bayesian Variable Selection in Generalized Linear Regression Model

摘要：

　　We describe the use of exact IBF sampling method for Bayesian variable selection in a generalized linear regression model.The exact IBF sampling method is a non-iterative sampling method which completely avoids the problem of convergence and slow convergence associated with iterative Markov China Monte Carlo (MCMC) methods.The idea is at first to utilize the sampling Inverse Bayesian Formula (IBF) to derive the conditional distribution of the identify vector given the observed data, and then to draw i.i.d sampling from the complete-data posterior distribution.Applications to simulated data sets suggest that our algorithms perform well in identifying relevant predictor variables.

关键词： Inverse Bayesian Formula Bayesian variable selection MCMC

作者: JIA Shuqin

作者单位: School of Mathematics, Shandong University, Jinan, P.R.China,250100

会议类型: 国际会议

会议名称: The 6th International Institute of Statistics & Management Engineering Symposium(IISMES2014) (第六届国际统计与管理工程学术研讨会)

会议地点: 山东 威海

会议语种:英文

页码: 902-907

在线出版日期: 2014-07-20（万方平台首次上网日期，不代表论文的发表时间）