报告题目：多准则最佳总体选取的经验贝叶斯方法 (Selecting the Best Population under Multiple Criteria: An Empirical Bayes Approach)主讲人概况：Wen-Tao Huang (黃文涛)博士/教授
Prof. Tamkang University,Taiwan
Editor-in-Chief, Inter. J. Inform.and Manag. Sci..
PhD. Department of Statistics,PurdueUniversity,West Lafayette,Indiana,USA, 1972
Institute of Statistical Sciences, Academia Sinica, Associate Fellow (1972-76), Research Fellow (1976-2000).
President ofTaiwanIntellectual Technology and Applied Statistics Association (June / 2006 -June /2008).
Associate Editor: Statistica Sinica (1991-93), J. App Math and Decision Sci ( 1997-2001)
J. Applied Math. and Comput.,
Communi. Statistics: Theory and Methods,
Statistics and Probability Letters,
Comput. Statistics & Data Analysis，et al.
There are more than 46 papers published including 25 SCI.
Consider k populations whose meanθ(i) and varianceσ(i) are all unknown. For given control valuesθ(0),σ(0) andδ(0), we are interested in selecting some population whose mean is the most close toθ(0) in the qualified subset in which each mean is no further thanδ(0) fromθ(0) and whose variance is less than or equal toσ(0). We focus on the normal populations though its method can be applied to other distributions. A Bayes approach is set up and an empirical Bayes procedure is proposed which has been shown to be asymptotically optimal. Some simulation study is also given.