High dimensional clustering for Covariance matrices

报告人：潘光明

报告地点：腾讯会议（会议ID：182 208 952）

报告时间：2020年10月22日星期四17:30-18:30

报告摘要：

This talk is about high dimensional cluster analysis for covariance matrices. We propose to use the Hadamard product so that the covariance matrices of the original vectors belonging to the different clusters could be transferred to the mean vectors of the new transferred vectors. In this way, after such a transformation, a kind of sample spiked covariance matrices appears so that its eigenvalue and eigenvector could be used to do clustering. Moreover, a type of adjusted U statistic has also been proposed to detect the significant entries of the population covariance matrices. Theoretic properties and extensive simulations have been developed for such a method.

主讲人简介：

潘光明，新加坡南洋理工大学教授，博士生导师。2005年博士毕业于中国科学技术大学统计金融系；之后在新加坡国立大学、台湾中山大学、荷兰埃因霍温科技大学做博士后和学术交流工作；自2008年以来，在新加坡南洋理工大学工作；2013年遴选为国际统计学会会员(Elected Member of International Statistical Institute)。研究领域包括计量经济理论、高维统计、随机矩阵、多元统计等。主持新加坡国家基金项目5项，已在《Journal of the Royal Statistical Society Series B》、《Annals of Statistics》、《Journal of the American Statistical Association》、《Annals of Probability》、《Annals of Applied Probability》、《Bernoulli》、《IEEE Transactions on Signal Processing》、《IEEE Transactions on Information Theory》等顶级统计学杂志上发表60余篇学术论文，担任《Random Matrices: Theory and Applications》杂志编委。