Bayesian Analysis of Semiparametric Mixed-Effects Models for Zero-Inflated Count Data
Bayesian Analysis of Semiparametric Mixed-Effects Models for Zero-Inflated Count Data
复制标题
DOI：
10.1080/03610920802468699
复制
发表时间：
2009-05
期刊：
影响因子：
--
通讯作者：
Cheng Xue-dong
中科院分区：
文献类型：
--
作者：
Cheng Xue-dong
近年来，零膨胀计数数据模型，例如零膨胀泊松（ZIP）模型，被广泛使用，因为在许多实际问题中，带有额外零的计数数据非常常见。为了对聚类或重复的相关计数数据进行建模，并以灵活的方式评估连续协变量或时间尺度的影响，考虑一类零膨胀计数数据的半参数混合效应模型。在本文中，我们基于数据增强方案为此类模型提出了完全贝叶斯推理，该方案反映了协变量的随机效应和零膨胀分布的混合。采用结合吉布斯采样器和 M-H 算法的计算高效的 MCMC 方法来获得模型参数的估计。最后，使用模拟研究和实际例子来说明所提出的方法。
In recent years, zero-inflated count data models, such as zero-inflated Poisson (ZIP) models, are widely used as the count data with extra zeros are very common in many practical problems. In order to model the correlated count data which are either clustered or repeated and to assess the effects of continuous covariates or of time scales in a flexible way, a class of semiparametric mixed-effects models for zero-inflated count data is considered. In this article, we propose a fully Bayesian inference for such models based on a data augmentation scheme that reflects both random effects of covariates and mixture of zero-inflated distribution. A computational efficient MCMC method which combines the Gibbs sampler and M-H algorithm is implemented to obtain the estimate of the model parameters. Finally, a simulation study and a real example are used to illustrate the proposed methodologies.