时间：2018年10月10（星期三）14：30-15：30

地点：学院南路校区，学术会堂606

报告题目：Model Confidence Bounds for Variable Selection

报告人：李扬教授，中国人民大学统计与数学学院

报告摘要：In this article, we introduce the concept of model confidence bounds (MCBs) for variableselection in the context of nested models. Similarly to the endpoints in the familiarconfidence interval for parameter estimation, the MCBs identify two nested models(upper and lower confidence bound models) containing the true model at a givenlevel of confidence. Instead of trusting a single selected model obtained froma given model selection method, the MCBs proposes a group of nested models ascandidates and the MCBs’ width and composition enable the practitioner toassess the overall model selection uncertainty. A new graphical tool — themodel uncertainty curve (MUC) — is introduced to visualize the variability ofmodel selection and to compare different model selection procedures. The MCBs methodology is implemented by a fast bootstrap algorithm that is shown to yieldthe correct asymptotic coverage under rather general conditions. Our MonteCarlo simulations and a real data example confirm the validity and illustratethe advantages of the proposed method.

报告人简介：李扬，中国人民大学统计学院教授、博士生导师，统计咨询研究中心主任，国际统计学会推选会员、国际生物统计学会中国分会青年理事委员、北京生物医学统计与数据管理研究会副秘书长。主要从事相关型数据分析, 模型选择与不确定性评价, 潜变量建模, 稳健统计, 临床试验设计等领域研究，承担国家重点研发计划精准医学研究重点专项子课题、国家自然科学基金面上项目等科研项目二十余项，发表Biometrics、Statistics in Medicine、Statistical Methods in Medical Research、统计研究、数理统计与管理等国内外期刊研究论文五十余篇。

本次活动受中央财经大学2018专题学术讲座项目资助。