A1556
Title: The asymptotic properties of the extreme eigenvectors of high-dimensional spiked models
Authors: Zhangni Pu - Northeast Normal University (China)
Xiaozhuo Zhang - Northeast Normal university (China) [presenting]
Abstract: Asymptotic behaviors of the extreme eigenvectors in the general spiked covariance matrix and sample canonical correlation matrix are investigated, where the dimension and sample size increase proportionally. The restrictive assumption of the block diagonal structure and the 4th moment in the population covariance matrix is eliminated. As for the SCC matrix, the asymptotic distribution of inner product of population and sample spiked eigenvectors is calculated. Simulation results are also presented to validate the key theoretical findings.
