HOME   /   Content


Guangming Pan: Can we trust non-stationary PCA?
time: Dec 26, 2019

Time:December 18, 2019 16:00-16:40(Wednesday)

Location:Conference room 1008, building B,Feicuihu Campusscience and education

Speaker:ProfessorGuangming Pan

From:Nanyang University of technology, Singapore

Organizer:School of Economics

Lecturer introduce:Guangming Pan, professor and doctoral supervisor of Nanyang University of technology, Singapore. In 2005, he graduated from the Department of statistics and finance of University of science and technology of China; later, he worked in the National University of Singapore, Sun Yat sen University of Taiwan and Eindhoven University of science and technology of the Netherlands for postdoctoral and academic exchange; since 2008, he worked in Nanyang University of Technology of Singapore; in 2013, he was selected as a member of the International Statistical Institute. Research fields include econometric theory, high-dimensional statistics, random matrix, multivariate statistics, etc. He has presided over 5 projects of Singapore National Fund, and has been listed in 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 inf More than 60 academic papers have been published in the top statistical magazines such as information theory, and served as the editorial board of random materials: Theory and applications.

Description:We examine some high dimensional data including morality data which are nonstationary and temporary correlated . The aim is to understand how time dependence and temporary correlation affect PCA. To this end we investigate the largest sample eigenvalues of separable covariance matrices. A new statistic is proposed to distinguish nonstationary factor models from unit root models.