带不可忽略缺失数据的再生散度随机效应模型的Bayes估计

摘要

在响应变量带有不可忽略缺失数据的前提下得到非线性再生散度随机效应模型的Bayes方法.缺失数据机制由Logistic回归模型定义,根据Gibbs抽样技术和MH算法得到模型参数、随机效应因子以及缺失数据机制中回归系数的联合Bayes估计,并进行了实例分析.%Bayesian method is considered to analyze nonlinear reproductive dispersion mixed models in which the response variables are missing with nonignorable missingness mechanism. The missingness is specified by a logistic regression model. Gibbs sampler and the MH is used to obtain the joint Bayesian estimates of parameters. The proposed procedure is illustrated by a real example.