Approximate Bayes factors and accounting for model uncertainty in generalised linear models
Approximate Bayes factors and accounting for model uncertainty in generalised linear models
复制标题
DOI：
10.1093/biomet/83.2.251
复制
发表时间：
1996-06-01
期刊：
影响因子：
2.7
通讯作者：
Raftery, AE
中科院分区：
文献类型：
--
作者：
Raftery, AE
基于积分的拉普拉斯方法，给出了广义线性模型的近似贝叶斯因子的求解方法。我们提出了一个新的近似，它只使用标准的计算机程序的输出估计广义线性模型，这似乎是相当准确的。适当的先验的参考集的建议，既代表的情况下，没有太多的先验信息，并评估结果的灵敏度先验分布。该方法可用于色散参数未知、存在过色散时比较链路函数、比较误差分布和方差函数。该方法可用于实现贝叶斯方法来考虑模型的不确定性。我们描述了一个应用程序的控制因素的存在下，模型的不确定性是大的和重要的相对风险的推断。实现这些方法的软件可从StatLib免费获得。
Ways of obtaining approximate Bayes factors for generalised linear models are described, based on the Laplace method for integrals. We propose a new approximation which uses only the output of standard computer programs for estimating generalised Linear models; this appears to be quite accurate. A reference set of proper priors is suggested, both to represent the situation where there is not much prior information, and to assess the sensitivity of the results to the prior distribution. The methods can be used when the dispersion parameter is unknown, when there is overdispersion, to compare link functions, and to compare error distributions and variance functions. The methods can be used to implement the Bayesian approach to accounting for model uncertainty. We describe an application to inference about relative risks in the presence of control factors where model uncertainty is large and important. Software to implement the methods is available at no cost from StatLib.