报告题目：Random Threshold Driven Tail Dependence Measures with Application to Precipitation Data Analysis

时间：2015年6月10（周三）14：00 - 15：00

地点：学院南路校区，学术会堂604

报告人：Professor Zhengjun Zhang, Department of Statistics, University of Wisconsin

摘要：

Identification of tail dependence among observations is important and challenging, but remains an open problem, due to the fact that tail dependence is primarily captured by values above thresholds. By introducing random thresholds, this paper establishes an approximation theory between conditional tail probabilities. The paper studies the tail quotient correlation coefficients (TQCC) with random threshold values. The new random threshold driven TQCC can be used to test the null hypothesis of tail independence under which the TQCC test statistics are shown to follow a Chi-squared distribution under two general scenarios respectively. The TQCC is shown to be consistent under the alternative hypothesis of tail dependence under a general approximation of max-stable distribution. We apply TQCC to investigate tail dependencies of daily precipitation in continental US. Our results, in the perspective of tail dependence, reveal nonstationarity, spatial clusters, and tail dependence in the precipitations cross continental US.

报告人简介：

张 正军教授为永利集团304am登录“手拉手”项目特聘教授，威斯康星大学统计系教授、副主任；北卡罗来纳大学教堂山分校统计学博士，北京航空航天大学管理工程博士。 美国统计学会、数理统计学会等多个学会会员，曾获得University of North Carolina教学奖等多项奖励，2010年入选剑桥名人录。主持有10余项美国自然科学基金等科研课题；在JASA等顶级统计学期刊发表学术论文50 余篇。同时担任Journal of Business and Economic Statistics等多个国际著名统计学期刊的副主编。