报告题目：Inherent Supervised Clustering with Low Rank
报 告 人：佘轶原 教授 佛罗里达州立大学
报告时间：2020年6月22日 9:00-10:00
报告地点：腾讯会议ID：808 349 325
或点击链接直接加入会议：
https://meeting.tencent.com/s/PavDBvmV3tx0
校内联系人：程建华 chengjh@jlu.edu.cn
报告摘要：
Modern high-dimensional methods commonly adopt the "bet on sparsity" principle while in the big-data era statisticians often face "dense" problems with large numbers of unknowns. This paper gives a mathematical formulation of low-rank supervised clustering to automatically group the predictors in building a multivariate predictive model. By use of linearization and block coordinate descent, a simple-to-implement algorithm is developed, which performs subspace learning and clustering iteratively with guaranteed convergence. We show a tight error bound of the proposed method, study its minimax optimality, and propose a new information criterion for parameter tuning, all with distinctive rates from the large body of literature based on sparsity. Extensive simulations and real-data experiments demonstrate the excellent performance of rank-constrained inherent clustering.
报告人简介：
佘轶原，佛罗里达州立大学统计系教授，2008年毕业于斯坦福大学，获得统计学博士学位。佘教授的主要研究方向包括：高维统计、统计机器学习、优化、信号处理、稳健统计和网络科学等领域，曾获得NSF CAREER Award，Florida State University Developing Scholar Award 等奖项，并先后担任Metrika，IEEE Transactions on Network Science and Engineering以及Journal of the American Statistical Association等顶级杂志的编委。
