学术预告-Arc transitive embeddings of graphsSemiparametric Bayesian analysis of accelerated failure time models with cluster structures

信息来源: 发布时间2017-7-14 14:54:49

 讲座主题:Semiparametric Bayesian analysis of accelerated failure time models with cluster structures

专家姓名:沈俊山

工作单位:首都经贸大学

讲座时间:201772810:10

讲座地点:数学院大会议室

主办单位:烟台大学数学与信息科学学院

内容摘要:

    In this talk, we develop a Bayesian semiparametric AFT model for survival data with cluster structures (BSP-DRM). We show through both simulation studies and analysis of Mayo clinic trial in PBC that the information pooling can significantly improve the efficiency of estimating regression coefficients in the AFT models. Moreover, the flexible accommodation of distributional heterogeneity greatly reduces potential estimation biases, and also improves estimation efficiency when the distributions of different clusters have different shapes.

主讲人介绍:

    兰州大学硕士、北京大学博士;主要从事生存分析、不完全数据分析、经验似然、 半参数模型推断、 Bayes 统计学等方面的研究. J. Amer. Statist. Assoc.Comput. Statist.Data Anal.Statist. Papers Ann. Inst. Statist. Math. J. Multivariate Anal.Statist. Probab. Lett.等国际著名学术刊物发表论文16. 主持完成了国家自然科学基金青年基金项目和博士后基金项目.