发布日期:2015-06-23点击: 发布人:统计与数学学院
主题: Recent advances for a class of new multivariate zero inflated/truncated/adjusted Poisson and related models with applications in insurance and other fields
主讲人:田国梁博士(香港大学统计与精算学系副教授)
时间:2015年7月4日(周六)下午14:30-16:00
地点:北院卓远楼305
主办单位:统计与数学学院
摘要:Count data with adjusted zeros and without zero-category arise quite frequently in studies of insurance, medicine, epidemiology, public health, biology, agriculture and other fields such as sociology, psychology, ecology and engineering. In this talk, I will introduce a class of newly developed multivariate zero inflated / truncated / adjusted Poisson (MZIP, MZTP, MZAP) and related distributions (e.g., Type I, II, III and IV MZIP, MZTP, MZAP distributions) by the author and his postgraduates/collaborators. First, I introduce the Type I MZIP distribution to model correlated multivariate count data with extra zeros. The distributional properties and efficient statistical methods are developed. Next, I will briefly introduce Type I MZAP models (including Type I MZIP and MZTP as special cases) by accounting for both inflation and deflation at zero. Third, to address the issues of over-dispersion and under-dispersion, other developments are highlighted by replacing Poisson with negative-binomial and generalized Poisson distributions.
田国梁博士简介:统计学博士,香港大学统计与精算学系副教授,研究领域包括零过多计数数据分析、不完全分类数据分析、敏感性问题抽样调查等,在国际主流学术刊物上共发表80 多篇统计学、医学统计方面的学术论文,在国际著名出版社Chapman and Hall/CRC3和John Wiley & Sons出版专著3部。目前担任Computational Statistics and Data Analysis, Statistics and Its Interface; Communications in Statistics - Theory and Methods; Communications in Statistics - Simulation and Computation等4个国际SCI期刊副主编。