
经济与金融系学术讨论会第208期
题目:Recent advances in causal inference
演讲人:谭若虚(数学科学学院 助理教授)
时间:2024年4月9日(周二)12:00-13:00
地点:同济大厦A楼505室
同步#腾讯会议:291301690
会议密码:517448
链接:https://meeting.tencent.com/dm/0isw150DHucn
摘要: Causal inference has gained extensive research interests in statistics, economics and computer sciences due to its numerous applications. The talk starts from a brief introduction to the Neyman-Rubin potential outcome model, the cornerstone of modern causal inference, followed by commonly-used estimation strategies of the average treatment effect (ATE). In particular, our work focuses on estimation of the ATE when the treatment variable is functional, a much more complicated case. We develop the so-called functional stabilized weight and three estimators for the ATE. We illustrate our method by an EEG dataset.
简介: 谭若虚博士,博士毕业于澳大利亚墨尔本大学,导师为Aurore Delaigle教授。他曾在香港大学统计与精算学院从事了两年博士后的研究,合作导师为尹国圣教授,现为同济大学数学科学学院的助理教授。谭博士的研究兴趣包括:函数型数据分析、流形学习、因果推断等,他的研究成果发表在《Statistica Sinica》、《Statistics in Medicine》等统计学期刊上。