概率统计| On multilinear PCA and Kronecker envelope PCA

题目：On multilinear PCA and Kronecker envelope PCA 

报告人：Professor Su-Yun Huang  (Institute of Statistical Science, Academia Sinica,Taiwan ) 

时间：2019年12月23日（周一）16: 00  

地点：浙江大学玉泉校区邵逸夫工商管理楼200-9 

摘要：PCA is one of the most popular dimension reduction methods. Many modern data analysis tools which target on large data set will adopt PCA as a built-in preprocessing step to reduce data dimensionality. Multilinear principal component analysis (MPCA) is an extension of PCA to tensor (or array) data. It preserves the natural Kronecker product structure of observations when searching for principal components. The main advantage of preserving the Kronecker product structure is the parsimonious usage of parameters in specifying the principal component subspaces, which mitigates the adverse influence of high-dimensionality, and hence, leads to efficiency gain in estimation and prediction. Note that PCA will convert possibly correlated variables to uncorrelated ones. However, it is not the case for MPCA. In some applications, decorrelation is necessary. Hence, the Kronecker envelope PCA is introduced. Some theoretical investigation will be provided. 

报告人简介：陈素云博士目前为台湾中央研究院统计科学研究所研究员。她本科和硕士毕业于台湾大学数学系(1983,1985)，1990年获美国普渡大学统计学博士学位。1990年至今一直在台湾中央研究院工作。研究领域包括: Statistical Machine Learning with Applications, Dimension Reduction, Robust Statistical Inference, Statistical Analysis for Biological Image Data, Nonparametric Inference. (http://www.stat.sinica.edu.tw/syhuang/) 
 

欢迎大家参加！    

联系人：庞天晓(txpang@zju.edu.cn)    

浙江大学数学科学学院统计研究所