Home > Academic Announcements > (Dec. 12) Optimal Convergence Rates of Canonical Correlation for Stochastic Process

(Dec. 12) Optimal Convergence Rates of Canonical Correlation for Stochastic Process

Last updated :2018-12-07

Topic: Optimal Convergence Rates of Canonical Correlation for Stochastic Process
Speaker: Prof. CHEN Dirong
(Beihang University)
Time: 10:00-12:00, Thursday, December 13, 2018
Venue: Room 415, Mathematics Building, Guangzhou South Campus, SYSU

Abstract:
Functional canonical correlation analysis (FCCA) has been used in many applications. Different from multivariate case, where the canonical vectors are obtained first and then the canonical variables, the canonical functions for functional data might not exist. Nevertheless, one can study canonical variables directly in a more general setting. The theoretical analysis of FCCA is more involved. In this talk we discuss the optimal estimations of canonical variables in FCCA.