报告题目：数据驱动的亚组分析

报告人：何叶 博士 可视化计算与虚拟现实四川省重点实验室

报告时间：5月29日上午10：00-10：30

报告地点：四川师范大学狮子山校区（锦江区静安路5号）第七教学楼C区7楼703

报告摘要：The existing methods for subgroup analysis can be roughly divided into two categories: finite mixture models (FMM) and regularization methods with an l_1-type penalty. In this paper, by introducing the group centers and l_2-type penalty in the loss function, we propose a novel center-augmented regularization(CAR) method; this method can be regarded as a unification of the regularization method and FMM and hence exhibits higher efficiency and robustness and simpler computations than the existing methods. In particular, its computational complexity is reduced from the O(n^2) of the conventional pairwise-penalty method to only O(nK), where n is the sample size and K is the number of subgroups. The asymptotic normality of CAR is established, and the convergence of the algorithm is proven. CAR is applied to a dataset from a multicenter clinical trial, Buprenorphine in the Treatment of Opiate Dependence; a larger R^2 is produced and three additional significant variables are identified compared to those of the existing methods.

专家简介：何叶，博士，四川师范大学，硕士生导师，毕业于西南财经大学，电子科技大学博士后，中国现场统计研究会资源与环境统计分会理事，研究方向为生存数据分析，个性化治疗，亚组分析，现已在《Biometrics》，《Statistics in Medicine》等国际期刊发表学术论文若干篇，主持国家自然科学基金青年项目，中国博士后科学基金面上项目等多个项目。