往期回顾

第三届中国人民大学国际统计论坛

2008-06-23

2008统计学国际论坛

会议手册

(2008年6月21-22日,北京•中国)

主办:中国人民大学统计学院

中国人民大学应用统计科学研究中心

国家统计局统计科学研究所

目   录

1. 学术委员会名单  2

2. 组委会名单  4

3. 会议日程表   5

4. 与会者报告摘要(按报告顺序)

(1)逸夫第一报告厅   17

(2)分会场一   21

(3)分会场二   37

(4)分会场三   52

(5)分会场四   69

(6)分会场五   86

(7)其他来稿   101

5.  附件

中国人民大学校园布局图   116



2008统计学国际论坛学术委员会

主席:

谢伏瞻  (国家统计局)

袁卫    (中国人民大学)


委员:(以拼音字母为序)

陈敏    (中国科学院系统与数学研究院)

范剑青  (普林斯顿大学)

冯士雍  (中国科学院系统与数学研究院)

巩馥洲  (中国科学院系统与数学研究院)

贺铿    (九三学社)

何书元  (北京大学)

何旭铭  (伊利诺斯大学厄巴纳-香宾分校 )

金勇进  (中国人民大学)

李冰    (宾夕法尼亚州立大学)

李纲    (国家统计局)

李强    (国家统计局)

邱东    (中央财经大学)

任若恩  (北京航空航天大学)

邵军    (威斯康星大学)

史宁中  (东北师范大学)

吴喜之  (中国人民大学)

肖红叶 (天津财经大学)

谢邦昌 (台湾辅仁大学)

许宪春 (国家统计局)

郁彬   (加州大学伯克利分校)

曾五一 (厦门大学)

张南   (广岛修道大学)

赵民德 (台湾中研院)

赵彦云 (中国人民大学)

朱力行 (香港浸会大学)



2008年统计学国际论坛会议组委员


主席:

袁卫      (中国人民大学)

副主席:

金勇进    (中国人民大学)

康君      (国家统计局)

委员:

赵彦云    (中国人民大学)

王晓军    (中国人民大学)

张波      (中国人民大学)

高敏雪    (中国人民大学)

宋大我    (中国人民大学)

许亦频    (国家统计局)

孙学光    (国家统计局)



2008统计学国际论坛日程表

Schedule of International Statistics Forum 2008

6月21日  (逸夫第一报告厅)

8:00-9:00报到

9:00-9:20 开幕式   (主持人:金勇进)

9:20-9:30 合影

9:30-9:50 茶歇

主持人:李纲

9:50-10:30 许宪春 国家统计局

报告题目:中国政府统计面临的挑战和发展机遇

10:30-11:10 贺铿 九三学社

报告题目:宏观经济中的统计特征数分析

11:10-11:50 袁卫 中国人民大学

报告题目:国内外统计高等教育的几个问题

12:00 -13:30 午餐 地点:中区食堂三层餐厅

主持人:朱力行

13:30-14:10 赵民德 台湾中央研究院统计研究所

报告题目:论惜墨

14:10-14:50 邵军 University of Wisconsin

报告题目: A unified theory for GMM Estimation in Panel Data Models with Measurement Error

14:50-15:30 蔡建文 University of North Carolina

报告题目: Recent development of Statistical methods for multivariate failure time data

15:30-15:50 茶歇


主持人:赵彦云

15:50:16:30 张南 日本广岛修道大学

报告题目:东亚地区资金循环分析模型及模拟分析

16:30-17:10 王武保 美国默克制药公司生物统计和研究决策科学部

报告题目: Applications of Biostatistics in the Pharmaceutical Industry


17:30--  欢迎晚宴  地点:中区食堂三层餐厅

6月22日 分会场1(明德主楼0404教室)

主持人:张波   主持人协助:田茂再

8:10-8:30 李冰,Department of Statistics ,Pennsylvania State University

报告题目:On principal components and regression: A statistical explanation of a natural phenomenon


8:30-8:50 朱力行,香港浸会大学数学系

报告题目:A Binary Response Transformation-Expectation Estimation in Dimension Reduction


8:50-9:10   耿直,北京大学数学科学学院概率统计系

报告题目:Learning a neighbor structure of a target in a causal network


9:10-9:30 石磊,云南财经大学统计与数学学院

报告题目:Case deletion, replacement and mean-shift model for diagnostics in linear mixed model


9:30-9:50 mustafa mohamed abdalla,economics faculty of economic and social studies university of Khartoum

报告题目:Measurement of Real Exchange Rate Misalignment in China 1978-2007

9:50-10:10 茶歇(明德主楼4层)

主持人:石磊   主持人协助:田茂再


10:10-10:30 Zhengjun Zhang,University of Wisconsin

报告题目:QUOTIENT CORRELATION: A SAMPLE BASED ALTERNATIVE TO PEARSON'S CORRELATION


10:30-10:50 Xiangrong Yin,Department of Statistics, University of Georgia

报告题目:A general adaptive L2 penalty method


10:50-11:10 闫军,Department of Statistics University of Connecticut

报告题目:Tests of serial independence for multivariate time series based on a M¨obius decomposition of the independence empirical copula process


11:10-11:30 赵国庆,中国人民大学经济学院

报告题目:非平稳时间序列的Granger因果检验研究


11:30-11:50 田茂再,中国人民大学统计学院

报告题目:Saddlepoint Approximations of Value-at-Risk


11:50-13:30 午餐    地点:中区食堂三层餐厅

主持人:闫军  主持人协助:金阳


13:30-13:45 赵作权,中国科学院科技政策与管理科学研究所

报告题目:空间分布统计:对统计学的挑战


13:45-14:00 戴平生; 陈建宝,厦门大学计划统计学院

报告题目:空间统计学研究应用综述


14:00-14:15 张军舰,北京工业大学应用数理学院

报告题目:一类新的积分型拟合优度检验


14:15-14:30 田玉斌,北京理工大学 理学院

报告题目:优化敏感性序贯试验设计


14:30-14:45 刘中华,中国人民大学统计学院

报告题目:Time-varying Doubly Weighted Estimation for the Additive Hazard Models in Case-Cohort Studies


14:45-15:00 栾文英,山东经济学院

报告题目:PoMix抽样的几个问题


15:00-15:30 茶歇(明德主楼4层)


15:30-17:00 戴世光教授百年诞辰学术研讨会

地点:明德主楼1016会议室

主持人:王晓军


6月22号 分会场2 (明德主楼0405教室)

主持人:郑纪伦     主持人协助:吕晓玲

8:10-8:30 Chunming Zhang,Department of Statistics, University of Wisconsin, Madison

报告题目:New Aspects of Bregman Divergence in Regression and Classification with Parametric and Nonparametric Estimation


8:30-8:50 荆柄义,香港科技大学数学系

报告题目:ANOVA for Jump-diffusion processes


8:50-9:10 谢邦昌,辅仁大学统计资讯学系

报告题目:Text Mining 及 Data Mining之实务与应用


9:10-9:30 王学仁,云南大学数学与统计学院统计系

报告题目:纸型识别的数学模型―Bayes多组


9:30-9:50 郑纪伦,中央研究院统计科学研究所

报告题目:On the Polynomial Berkson Model

9:50-10:10

茶歇(明德主楼4层)


主持人:王学仁     主持人协助:吕晓玲

10:10-10:30 刘金山,华南农业大学理学院统计系

报告题目:Linear sufficiency in a general growth curve model and some changes due to transformation


10:30-10:50 吕晓玲,中国人民大学统计学院

报告题目:A New Two-Group Forecast and Selection Method for Direct Mail: Based on the Repeat Buying Theory


10:50-11:10 李竹渝,四川大学数学学院

报告题目:金融时间序列的模糊结构初探


11:10-11:30 王华;金勇进,厦门大学  中国人民大学统计学院

报告题目:统计数据准确性评估的误差效应分析方法



11:30-13:30

午餐     地点:中区食堂三层餐厅


主持人:王华     主持人协助:蒋妍

13:30-13:45 唐启义,浙江大学农业与生物技术学院

报告题目:常用数据分析技术分类初探


13:45-14:00 林海明,广东商学院经济贸易与统计学院

报告题目:初始因子应用中六个问题的解析


14:00-14:15 聂巧平 冯蕾,天津商业大学经济学院;国家统计局统计研究所

报告题目:考虑结构突变的单位根检验程序研究


14:15-14:30 侯志强,北方工业大学

报告题目:中国劳动力调查样本轮换方法研究


14:30-14:45 戴平生; 曾五一,厦门大学计划统计学院

报告题目:用修正四联表分析我国各地区高等教育效率


14:45-15:00 孙慧钧,东北财经大学统计学院

报告题目:市场调查中问卷设计改进之探讨



15:00-15:30 茶歇(明德主楼4层)


15:30-17:00 戴世光教授百年诞辰学术研讨会

地点:明德主楼1016会议室

主持人:王晓军


6月22号 分会场3 (明德主楼0406教室)


主持人 :石庆焱       主持人协助:王燕

8:10-8:30 Huiyu Zhang,Reserving Actuary, Cuna Mutual Insurance Group

报告题目:Two Simulation Approaches for Reserve Variability Estimates


8:30-8:50 刘乐平,天津财经大学 统计学院

报告题目:Two-stage Adaptive Risk Control Procedure in Loss Reserving


8:50-9:10 孟生旺,中国人民大学统计学院

报告题目:混合泊松回归与索赔频率估计


9:10-9:30 张连增,南开大学风险管理与保险学系

报告题目:未决赔款准备金评估的对数正态模型及其WinBUGS实现


9:30-9:50 王传玉; 申文康,安徽工程科技学院 应用数理系

报告题目:一类延迟更新风险模型破产概率的等价关系,两类NCD系统在随机优序下的关系


9:50-10:10 茶歇(明德主楼4层)


主持人 :刘乐平       主持人协助:王燕

10:10-10:30 黄向阳,中国人民大学统计学院

报告题目:形成中的数据质量观及其影响


10:30-10:50 王晓军 钱珍,中国人民大学统计学院

报告题目:中国城乡统一的国民年金制度探讨


10:50-11:10 石庆焱,国家统计局统计科学研究所

报告题目:个人信用评分模型的一个扩展


11:10-11:30 柳会珍,北京化工大学

报告题目:股市波动的长记忆过程研究


11:30-11:50 韩海波,兰州商学院统计学院

报告题目:金融风险定义与测量的新探索——基于小波分析的波动频率时变性分析



11:50-13:30

午餐         地点:中区食堂三层餐厅


主持人 :孟生旺      主持人协助:肖宇谷

13:30-13:45 张建标 王传玉,安徽工程科技学院应用数理系

报告题目:两类NCD系统在随机优序下的关系

13:45-14:00 刘潭秋;孙湘海,长沙理工大学经济与管理学院

报告题目:中国股票市场收益的非线性行为研究


14:00-14:15 陈昊,江苏电力公司南京供电公司

报告题目:A Comparative Study on Load Forecasting Via Asymmetric ARCH Models


14:15-14:30 蔡正高,王晓军,中国人民大学统计学院

报告题目:Applicability of hedging longevity risks in China


14:30-14:45 王博,中国人民大学信息学院

报告题目:广义复合非齐次双Poisson过程的破产概率


14:45-15:00 尹伟,南开大学风险管理与保险学系

报告题目:用MatLab实现离散盈余过程有限时间内破产概率的计算


15:00-15:15 黄媛,中国人民大学统计学院

报告题目:Wilkie精算随机模型在中国的应用


15:15-15:30 茶歇(明德主楼4层)


15:30-17;00 戴世光教授百年诞辰学术研讨会

地点:明德主楼1016会议室

主持人:王晓军


6月22号 分会场4 (明德主楼0407教室)


主持人 :吴喜之       主持人协助:王渝

8:10-8:30 Haibo Zhou,the School of Public Health, the Center for Environmental Medicine, Asthma and Lung Biology

报告题目:Outcome-Dependent Sampling and Its Application to

Environmental Health Studies


8:30-8:50 刘韫宁,首都医科大学公共卫生与家庭医学学院

报告题目:基于灰度共生矩阵和小波变换提取CT图像肺小结节纹理建立人工神经网络预测模型


8:50-9:10 施红英,温州医学院环境与公共卫生学院

报告题目:SAS、SPSS、MLwiN统计软件在数据层次结构判断中的应用


9:10-9:30 李赞华,北京中医药大学循证医学中心

报告题目:中医药治疗糖尿病肾病随机对照试验文献中常见统计学问题


9:30-9:50 郭剑,津市塘沽区疾病预防与控制中心

报告题目:多水平模型在高血压病影响因素研究中的应用


9:50-10:10 王媛,天津医科大学卫生统计学教研室

报告题目:天津市社区居民高血压影响因素结构方程模型分析


10:10-10:30 茶歇 (明德主楼4层)

主持人 :高敏雪      主持人协助:薛薇


10:30-10:50 林筱文,福州大学管理学院统计系

报告题目:官方数据应用:通货膨胀预警系统研究,基于先行指标的省级宏观经济预测系统


10:50-11:10 斯琴,内蒙古财经学院统计系

报告题目:灰色系统预测方法及应用


11:10-11:30 季迎春; 陶磊,云南省思茅财经学校

报告题目:构建我国少数民族体育产业统计指标体系的若干思考


11:30-11:50 侯瑜,东北财经大学 经济与社会发展研究院

报告题目:污染事件的人体健康损失核算:方法与实践


11:50-13:30  午餐  地点:中区食堂三层餐厅


主持人 :林筱文      主持人协助:王星


13:30-13:45 郭志伟,东北财经大学统计学院

报告题目:北京市土地资源承载力综合评价研究


13:45-14:00 胡恩生,华侨大学商学院经济系

报告题目:新时期农民参与新型农村合作医疗的意愿及其决定因素的实证分析--基于泉州试点县、区的调查


14:00-14:15 王吉培,.西南财经大学统计学院

报告题目:基于IGARCH投影寻踪回归的国际油价走势拟合模型


14:15-14:30 许晓娟,中国人民大学统计学院

报告题目:中国外资经济独资化对技术溢出的影响


14:30-14:45 顾玲,河北大学经济学院

报告题目:我国城乡收入差距与其影响因素的典型相关分析


14:45-15:00 王丹丹,中国人民大学 统计学院

报告题目:中国国际服务贸易统计改进及其影响评估


15:00-15:15 刘崇光,中国人民大学统计学院

报告题目:企业寿命生命表的编制初探


15:15-15:30 茶歇(明德主楼4层)


15:30-17;00 戴世光教授百年诞辰学术研讨会

地点:明德主楼1016会议室

主持人:王晓军


6月22号 分会场5 (明德主楼0408教室)

主持人 :李金昌       主持人协助:李静萍

8:10-8:30 余芳东,国家统计局

报告题目:关于世界银行推算中国购买力平价的方法、结果及问题研究


8:30-8:50 白先春,南京财经大学统计系

报告题目:城市可持续发展中另类变量的选择


8:50-9:10 范秀荣,贺本岚,重庆工商大学数学与统计学院

报告题目:人力资源统计学的思考


9:10-9:30 杨京英,国家统计局统计科学研究所

报告题目:信息消费系数与信息化发展阶段研究报告


9:30-9:50 罗乐勤,吴燕华,厦门大学计划统计系

报告题目:住户无付酬服务时间使用的调查与分析


9:50-10:10 茶歇(明德主楼4层)


主持人 :杨京英       主持人协助:李静萍


10:10-10:30 刘保珺,关欣,陈平,天津财经大学;中国工商银行北京分行

报告题目:区域间产业结构差异成因分析模型及其应用


10:30-10:50 汪彩玲,陈相成,厦门大学经济学院,

报告题目:区域经济差异及其适度区间的确定


10:50-11:10 李武选,长安大学经济与管理学院

报告题目:再谈分段拟合建模技术及其应用


11:10-11:30 李金昌,浙江工商大学

报告题目:略论统计思想研究

11:50-13:30 

午餐 地点:中区食堂三层餐厅

主持人 :范秀荣      主持人协助:肖争艳


13:30-13:45 孔建新,暨南大学工业工程管理在职研究生

报告题目:单峰分布在社会经济应用的探讨,改进单侧规范过程能力指数计算公式的研究


13:45-14:00 张宝军,中央财经大学统计学院

报告题目:经济增长的部门均衡性研究


14:00-14:15 谢润椿,暨南大学经济学院统计学系

报告题目:广东省各市旅游业发展水平比较分析


14:15-14:30 何强,中央财经大学统计学院

报告题目:1993年SNA修订对中国官方核算的影响


14:30-14:45 朱婷,天津财经大学数量经济学

报告题目:关于京津区域的区域经济问题的实证研究


15:45-15:30 茶歇(明德主楼4层)

15:30-17;00 戴世光教授百年诞辰学术研讨会

地点:明德主楼1016会议室

主持人:王晓军

论文摘要

論惜墨

趙民德

(中央研究院統計科學研究所,台北南港)


當我讀書的時候, 指導老師教我們不要亂寫論文. 那時候,一個世界知名的大教授, 也不過二三十篇論文, 也沒有人認為他們學問不夠, 仍然到處受到尊敬。


到了今天, 這樣的情況有了極大的變化: 發表百篇論文以上的學者, 在各領域裡比比皆是; 即使在一般論文數較少的統計專業也屢見不鮮。


當然我們都知近道原因。但是這對我們是好是呸坯壞? 對研究是好是壞? 本文不算是學術報告, 因為我沒有預備講我做了甚麼; 本文只是反省, 說的是有些事我們也許不應該拚命努力去做。

A Unified Theory for GMM Estimation in Panel Data

Models with Measurement Error

Jun Shao

Abstract:Panel data allow correction for measurement error without using external information. Griliches and Hausman (1986) proposed using the generalized method of moments (GMM) or optimal weighting to efficiently combine instrumental variable (IV) estimators. Wansbeek (2001) applied GMM based on moment conditions expressed in the form of the Kronecker product. This paper studies some issues that are not addressed or not fully addressed by Griliches and Hausman (1986) or Wansbeek (2001) but are crucial to applications of these two approaches, these include how to choose instruments, what is the optimal weighted IV estimator, how to explicitly construct GMM estimators, how to remove the redundancy of the moment conditions constructed by Wansbeek (2001), and the existence of optimal GMM estimators. We unify Griliches and Hausman's and Wansbeek's approaches by establishing their equivalence. We also consider models with exogenous regressors and models with nonclassical assumptions. We apply the methods in this paper to revisit a investment controversy, viz., whether financially constrained firms respond to internal funds such as cash flow more sensitively than financially unconstrained firms.

Statistical methods for multivariate failure time data

Jianwen Cai

(Department of Biostatistics, University of North Carolina)

Abstract: Multivariate failure time data arise in many contexts. Some examples are: in epidemiological cohort studies in which the ages of disease occurrence are recorded for members of families; in animal experiments where treatments are applied to samples of littermates; in clinical trials in which individual study subjects are followed for the occurrence of multiple events; or, in intervention trials involving group randomization.

A common feature of the data in these examples is that the failure times might be correlated.

Over the past few years, we have developed statistical methods for various practical situations involving correlated failure time data.

In this talk, I will give an overview of these developments. I will focus on three general areas: (1) semiparametric methods for making inferences on linear and nonlinear covariates effects;(2) variable selection; and (3) joint modeling of recurrent event and recurrent marker. Examples will be drawn from biomedical studies

The Theory and Practice on Monetary and Financial Statistics

-A case survey in Zambia -

Nan Zhang

Hiroshima Shudo University (Japan)

The Monetary and Financial Statistics Manual (MFSM) was made by International Monetary Fund (IMF) in 2000, and it is to offer guidelines for the presentation of monetary and financial statistics. In addition to their role in assisting in monetary policy formulation and monitoring, the statistics covered in this volume form a basis for the development of a statistical framework for assessing financial sector. And the MFSM is harmonized with the 1993 SNA. Because the MFSM focuses on stocks and flows for financial corporation’s sector, it may, for the most part, be seen as extending and elaborating on the 1993 SNA in this area.

This report documented the principal findings and recommendations of the monetary and financial statistics mission from the Statistics Department (STA) of the IMF that visited Lusaka 2008. The main objectives of the mission were to: (i) review progress made by the Bank of Zambia (BOZ) in implementing the recommendations of the April/May 2005 monetary and financial statistics mission; (ii) assist in expanding the institutional coverage of the depository corporations survey; (iii) initiate work on the collection of data from other financial corporations and compilation of a financial corporation survey; (iv) verify the components of measures of the money supply and compilation of these monetary aggregates; and (v) discuss the units of central government used in classifying the deposits of government agencies at commercial banks.

The mission found that progress has been achieved in many areas, mainly in (i) timeliness and periodicity the BOZ submitted to STA, on a regular monthly basis, incorporating the accounts of all nonbank deposit-taking institutions and the compilation of the Form 1SR for the central bank’s accounts. And (ii) the compilation of the Form 2SR for the other depository corporations’ accounts, based on the recording system. Holdings of government securities by commercial banks reported. On the other hand, the implementation of some recommendation hasn’t been done, and the implementation recommendation has been slow. One important recommendation that had not been adopted was about collecting infrastructural data for recording of commercial banks.

Applications of Biostatistics in the Pharmaceutical Industry

William Wubao Wang, Ph.D

(Department of Biostatistics and Research Decision Sciences (BARDS)-China Operation)

Merck Research Laboratories

Merck & Co, Inc

Abstract: The growing trend of global clinical drug development demands increasing number of biometrics-related professionals (biostatisticians, clinical database administrator and clinical programmers) across the globe.  This trend is most recently reflected in the establishment of research and development operations in China and India by world's major pharmaceutical companies.

In this talk, we will provide an overview of the application of statistics and the role of clinical statisticians in the development of drug/vaccine products. We will discuss some recent trend in clinical study designs and analyses.  The Rotavirus Efficacy and Safety Trial (REST, n~72,000) will be used to illustrate some of these considerations (Vesikari et al. 2006, NEJM, 354(1): 23-33).

On principal components and regression: A statistical explanation of a natural phenomenon

Bing Li

(Professor of Statistics, Department of Statistics, Pennsylvania State University)

(joint work with Andreas Artemiou)

Abstract: In this talk we give a probabilistic explanation of a phenomenon that is frequently observed but whose reason is not well understood. That is, in a regression setting, the response (Y) is often highly correlated with the leading principal components of the predictor (X) even though there seems no logical reason for this connection. This phenomenon has long been noticed and discussed in the literature, and  has received renewed interest recently because of the need for regressing Y on  X of very high dimension, often with comparatively few sampling units, in which case it seems natural to regress on the first few principal components of X.

This work stems from a discussion of a recent paper by Cook (2007) which, along with other developments, described a historical debate surrounding, and current interest in, this phenomenon.

Learning a neighbor structure of a target in a causal network

Zhi Geng, You Zhou, Changchang Wang and Jianxin Yin

(School of Mathematical Sciences, Peking University, Beijing 100871, China)

Abstract: In causal discovery, a key issue is to discover causes of a target feature of interest. From observational studies, it is difficult to discover causes and effects of the target from a large number of observed features and especially to distinguish causes from effects. Discovering causal structures and further distinguishing causes from effects of a target are useful not only for prediction in the cases with external interventions, but also they are valuable for studying causal mechanisms, making decision and evaluating treatment effects. It is not easy to discover the parent nodes of a target in a large causal network. Although there are many structural learning approaches for discovering a global network, it is well known that learning a global network in an NP-Hard problem. If we are only interested in the neighbor of a target, it is inefficient and unnecessary to learn a global network.

Given a target variable of interest, we try to discover only the neighbor structure of the target. We propose in this paper two partial orientation and local structural learning (POLSL) approaches, Local-Graph and PCD-by-PCD. In the POLSL approaches, we locally discovery the edges connected to the target and only try to orient these edges so that we can distinguish the parents from the children of the target. We can theoretically show that the approach can correctly obtain the edges connected to the target and their orientations. The POLSL approaches can greatly reduce computational complexity of structural learning, and their statistical test is more powerful than a global learning approach.

Case Deletion, Replacement and Mean-shift Model

for Diagnostics in Linear Mixed Models

Lei Shi

(Statistics and Mathematics School, Yunnan University of Finance and Economics, Kunming, 650221, PRC)

Gemai Chen

(Department of Mathematics and Statistics, University of Calgary, AB, T2N 1N4, Canada)

Abstract: Case deletion, replacement and mean-shift model are frequently used in the detection of influential observations and outliers. In the general linear model with known covariance matrix, it is known these three techniques lead to the same results for the update formulae of the regression coefficients estimation. However if the covariance matrix is indexed by some unknown parameters which have to be estimated from the data set, the situation will be different. This paper study this issue in linear mixed model in which the parameter estimation are obtained iteratively. We propose several identities of case deletion method to derive and simplify the diagnostics for the estimation of parameters in covariance matrix and investigate the equivalence and non equivalence among three techniques. A example show that the misuse of replacement or mean-shift model for deletion diagnostics will lead to incorrect results.

Keywords: Case deletion, replacement, mean-shift model, diagnostics, influential observations.

Measurement of Real Exchange Rate Misalignment in China 1978-2007

Mustafa mohamed abdalla (B.Sc., M.Sc., Ph.D.)

(assistant professor in economics department of economics

faculty of economic and social studies university of khartoum

p o box 321  tel: 794923 mobile: 0918206239)

Abstract: The research is empirical and analytical in nature, methods of behavioural equilibrium exchange rate as well as purchasing power parity PPP, Capital enhanced equilibrium exchange rate and fundamental equilibrium exchange rate will be maniplualte to measure China RMB over/undervaluation.


QUOTIENT CORRELATION: A SAMPLE BASED ALTERNATIVE TO PEARSON'S CORRELATION

Zhengjun Zhang,

(University of Wisconsin)


Abstract: The quotient correlation is defined here as an alternative to Pearson's correlation that is more intuitive and flexible in cases where the tail behavior of data is important. It measures nonlinear dependence where the regular correlation coefficient is generally not applicable. One of its most useful features is a test statistic that has high power when testing nonlinear dependence in cases where the Fisher's $Z$-transformation test may fail to reach a right conclusion. Unlike most asymptotic test statistics, which are either normal or $\chi2$, this test statistic has a limiting gamma distribution (henceforth, the gamma test statistic). More than the common usages of correlation, the quotient correlation can easily and intuitively be adjusted to values at tails. This adjustment generates two new concept -- the tail quotient correlation and the tail independence test statistics, which are also gamma statistics. Due to the fact that there is no analogue of the correlation coefficient in extreme value theory, and there does not exist an efficient tail independence test statistic, these two new concepts may open up a new field of study. In  addition, an alternative to Spearman's rank correlation, a rank based quotient correlation, is also defined. The advantages of using these new concepts are illustrated with simulated data and real data analysis of internet traffic and asset returns.


A general adaptive L2 penalty method

Xiangrong Yin*

(Department of Statistics, University of Georgia, 204 Statistics Building, Athens, GA 30602)


Abstract: In this talk, we propose a general adaptive L2 penalty method. We show that it is a unfied method of adaptive lasso, adaptive ridge and adaptive elastic net penalties, making it a data-driven approach. We suggest new algorithms, providing important alternatives to the respective existing algorithms. Simulations and data analysis confirm the usefulness o our new approach.


会议提示
  • 2018.04.15

    提交报告截止日期

  • 2018.04.30

    会议报名截止日期

  • 2018.04.30

    教师培训报名截止日期

联系我们
  • 联系人

    张育铭 陈   琳
  • 电话

    86-10-62511318
  • 邮箱

    ISF2018@ruc.edu.cn
  • 地址

    北京市海淀区中关村大街59号中国人民大学明德主楼10层