数学与统计学院学术报告[2020] 016号

(高水平大学建设系列报告369号)

报告题目: Subgroup analysis of linear models with measurement error

报告人：秦国友 教授（复旦大学）

报告时间：6月7日9:00-10:00

报告地点：腾讯会议 会议号572 205 008　　　　　　　　　　　　　　　　　　　　　　

报告内容：

Heterogeneity exists in population and people may benefit differently from the same treatment or services. Correctly identifying different subgroups according to their outcome such as treatment responses plays an important role in providing data based evidence for intervention in practice. With few discussions on subgroup analysis in regression with measurement error, we propose a new estimation method of subgroup analysis with the existence of measurement error under the linear regression model. First, we develop an objective function based on the unbiased estimating equations with two replicate measurements and the concave penalty with the pairwise differences of coefficients. The proposed method can identify the subgroups and estimate coefficients simultaneously with the consideration of measurement error. Second, we derive the algorithm with concave penalties based on the alternating direction method of multipliers and demonstrate its convergence. Third, we prove that the proposed estimates hold the consistency and asymptotic normality. The performance of the proposed method and its asymptotic property are evaluated by simulations. Finally, we apply our method to the data from the Lifestyle Education for Activity and Nutrition study and identify two subgroups with one group having significant treatment effect.

报告人简历：

秦国友，教授，博士生导师。公共卫生学院生物统计学教研室主任，主要从事生物统计学方法学和应用研究，包括针对复杂数据、复杂统计模型的统计方法研究，以及生物统计学方法在医学和公共卫生领域的应用。在Biometric, Biostatistics, Statistics in medicine, BMJ等生物统计和医学权威期刊上发表60余篇SCI论文。在纵向数据方面相关研究工作获得教育部高等学校科学研究优秀成果奖二等奖。

　　　　　　　　 　　　　　　　　数学与统计学院

　 2020年6月3日