泊松逆高斯回归模型的贝叶斯统计推断

Bayesian Statistical Inferences for Poisson Inverse Gaussian Regression Models

本文研究泊松逆高斯回归模型的贝叶斯统计推断.基于应用Gibbs抽样,Metropolis-Hastings算法以及Multiple-Try Metropolis算法等MCMC统计方法计算模型未知参数和潜变量的联合贝叶斯估计,并引入两个拟合优度统计量来评价提出的泊松逆高斯回归模型的合理性.若干模拟研究与一个实证分析说明方法的可行性.

Bayesian statistical inferences for the Poisson inverse Gaussian regression models are studied in this paper.MCMC techniques containing the Gibbs sampler and the Metropolis-Hastings algorithm as well as the Multiple-Try Metropolis algorithm are used to calculate joints the Bayesian estimations of unknown parameters and the latent variables,the two goodness-of-fit statistics for assessing the plausibility of the posited Poisson inverse Gaussian regression Models are introduced.The proposed methodology is demonstrated by using simulation studies and a real example.