Tree-based Survival Models and Precision Medicine

题目：Tree-basedSurvival Models and Precision Medicine

报告人： Yifan Cui (宾夕法尼亚大学)

时间：2018.12.19(周三)下午 3:00

地点：紫金港校区管理学院行政楼14楼1417报告厅

摘要：

In the first part, we develop a theoretical framework for survivaltree and forest models. We first investigate the method from the aspect ofsplitting rules. We show that existing approaches lead to a potentially biasedestimation of the within-node survival and cause non-optimal selection of thesplitting rules. Based on this observation, we develop an adaptiveconcentration bound result which quantifies the variance component for survivalforest models. Furthermore, we show with two specific examples how theseconcentration bounds, combined with properly designed splitting rules, yieldconsistency results. In the second part, we focus on one application ofsurvival trees in precision medicine which estimates individualized treatmentrules nonparametrically under right censoring. We extend the outcome weightedlearning to right censored data without requiring either inverse probability ofcensoring weighting or semi-parametric modeling of the censoring and failuretimes. To accomplish this, we take advantage of the tree-based approach tononparametrically impute the survival time in two different ways. In simulationstudies, our estimators demonstrate improved performance compared to existingmethods. We also illustrate the proposed method on a phase IIIclinical trial of non-small cell lung.

欢迎广大师生踊跃参加!

联系人：苏中根（suzhonggen@zju.edu.cn）