Bayesian Model Selection in High-Dimensional Settings.

Bayesian Model Selection in High-Dimensional Settings.
复制标题

高维设置中的贝叶斯模型选择。

DOI:
10.1080/01621459.2012.682536
复制
发表时间:
2012
影响因子:
3.7
通讯作者:
Rossell D
Rossell D
中科院分区:
数学1区
文献类型:
--
作者:
Johnson VE;Rossell D

文献摘要

参考文献

被引文献

相似文献

结合到贝叶斯模型选择过程中的标准假设导致了与常用的惩罚似然方法不具竞争力的过程。通过对模型参数施加非局部先验密度,我们提出了对这些方法的修改。我们证明了当可能的协变量的数目p被观测的数目n所限定时,所得到的模型选择过程在线性模型设置中是一致的,这一性质还没有扩展到其他模型选择过程。除了一致地识别真实模型之外,建议的程序还提供了对每个识别的模型正确的后验概率的准确估计。通过仿真研究,我们证明了这些模型选择方法在一定的仿真环境下与常用的惩罚似然方法具有相同或更好的性能。主要定理的证明在网上可获得的补充材料中提供。
Standard assumptions incorporated into Bayesian model selection procedures result in procedures that are not competitive with commonly used penalized likelihood methods. We propose modifications of these methods by imposing nonlocal prior densities on model parameters. We show that the resulting model selection procedures are consistent in linear model settings when the number of possible covariates p is bounded by the number of observations n, a property that has not been extended to other model selection procedures. In addition to consistently identifying the true model, the proposed procedures provide accurate estimates of the posterior probability that each identified model is correct. Through simulation studies, we demonstrate that these model selection procedures perform as well or better than commonly used penalized likelihood methods in a range of simulation settings. Proofs of the primary theorems are provided in the Supplementary Material that is available online.
DOI: 10.1214/08-aos606
发表时间: 2009-06-01
影响因子: 4.5
作者:
Casella, George;Giron, F. Javier;Moreno, Elias
通讯作者: Moreno, Elias
DOI: 10.1214/009053604000000256
发表时间: 2004-06-01
影响因子: 4.5
作者:
Fan, JQ;Peng, H
通讯作者: Peng, H
DOI: 10.1198/016214506000000735
发表时间: 2006-12-01
影响因子: 3.7
作者:
Zou, Hui
通讯作者: Zou, Hui
DOI: 10.1214/009053604000000067
发表时间: 2004-04-01
影响因子: 4.5
作者:
Efron, B;Hastie, T;Tibshirani, R
通讯作者: Tibshirani, R
DOI: 10.1111/j.1467-9868.2005.00503.x
发表时间: 2005-01-01
影响因子: 5.8
作者:
Zou, H;Hastie, T
通讯作者: Hastie, T