报告时间：2016年5月23日下午1:00—2:00
报告地点：统计楼105
报告题目: Subgroup Identification Based on Multiple Outcomes, Meta Data, and Longitudinal Data
报告人：Dr Menggang Yu, University of Wisconsin - Madison
摘要: We first consider a novel method for subgroup identification based on multiple outcomes. Our method is motivated by the situation when multiple outcomes share common influential features in subgroup identification and hence an integrated analysis of all outcomes would provide more power in variable selection and better performance in prediction, compared with separate analyses for each outcome. The proposed method is based on a classification framework in subgroup identification and multivariate boosting. We adapt the boosting to incorporate weights and extend it to various loss functions. Hence the updated algorithm is a highly flexible approach due to the possible choice of loss functions and base learners. The approach can deal with high dimension in both covariates and responses.