报告题目：From Model Selection to Model Averaging: A Comparison Studies for Nested Linear Models

报 告 人：张新雨 研究员 中科院系统所

报告时间：2021年11月5日 下午 15:00-16:00

报告地点：腾讯会议 ID：553 620 445

或点击链接直接加入会议https://meeting.tencent.com/dm/TdDOpyJrEmHR

校内联系人：赵世舜 zhaoss@jlu.edu.cn

报告摘要：Model selection (MS) and model averaging (MA) are two popular approaches for dealing with model uncertainty. Most existing literature is limited to the optimal properties of MS and MA in their own terms, not their comparison. A foundational issue is whether MA offer any significant improvement over MS. Recently, Peng and Yang (2021) has answered this question in the nested model setting with series expansion. In this paper, our goal is to answer the same question in a linear regression framework. We further broaden the scope of analysis to compare MAs with the weights come from three popular weight sets. Simulation studies support the theoretical findings in a variety of settings.

报告人简介：张新雨，中科院系统所研究员，Texas A&M大学博士后、Penn State 大学Research Fellow。主要研究方向为模型平均、模型选择、组合预测等。国家杰出青年科学基金获得者，主持3项国家自然科学基金，目前担任《JSSC》、《SADM》、《系统科学与数学》、《应用概率统计》编委和《Econometrics》客座主编。