报告时间：2022年11月16日，上午9：00-10：00
报告地点：腾讯会议（会议ID：269-614-862）
报告嘉宾：周迎春
报告主题：High Dimensional Covariate Selection in Causal Effect Estimation for Multivariate Continuous Treatments
报告摘要：
High Dimensional Covariate Selection in Causal Effect Estimation for Multivariate Continuous Treatments
Causal inference methods based on observational data generally estimate causal effects through balancing covariates using e.g. propensity score methods. However, previous studies have shown that incorrect selection of covariates in the propensity score model can lead to bias and efficiency loss. Therefore, it is essential to select appropriate covariates from high or even ultra-high dimensional covariates. In this talk, a new method called double screening prior adaptive lasso (DSPAL) is proposed to select confounders and predictors of the outcome, which combines the adaptive lasso method with the marginal conditional (in)dependence prior information to select target covariates. The DSPAL method is suitable for causal effect estimation for multivariate continuous treatments with high-dimensional or even ultra-high dimensional covariates. It can be used in the case of both parametric and nonparametric outcome models. Theoretical properties and simulation results indicate that DSPAL selects all confounders and predictors consistently and outperforms other methods under various scenarios. In real data analysis, the method is applied to estimate the causal effect of a three-dimensional continuous environmental treatment on cholesterol level and enlightening results are obtained.
个人简介：
周迎春, 华东师范大学统计学教授、博士生导师。复旦大学学士，美国波士顿大学（Boston University）统计学博士。主要研究函数型数据分析、因果推断以及生物统计中的各类问题。主持国家自然科学基金项目两项、参与国家自然科学基金重点项目两项，获上海市“浦江人才计划”，承担了多项医院与制药公司的企事业横向项目。在Annals of Applied Statistics, Statistics in Medicine等国际一流期刊发表论文30余篇。担任中国工程概率统计学会理事、中国现场统计研究会高维数据统计分会理事等。2017年获上海市教学成果一等奖，2018年获第三届全国应用统计专业学位研究生教育教学成果奖一等奖，2019年获宝钢教育基金优秀教师奖。
扫描下方二维码报名↘
所有消息会在两个群中同步通知
请大家不要重复加群~
