B0664
Title: High-dimensional tests for principal component analysis
Authors: Thomas Verdebout - Universite Libre de Bruxelles (Belgium) [presenting]
Davy Paindaveine - Universite libre de Bruxelles (Belgium)
Christine Cutting - Universite Libre de Bruxelles (Belgium)
Abstract: Principal Component Analysis (PCA) is one of the most important tools in multivariateanalysis. Nowadays, it is getting more and more popular in the statistical communitysince it is a specific high-dimensional problem. Indeed, the main objective of PCA isdimension reduction. In this paper, we tackle the problem of testing that the first principal component can be obtained by projecting the data along a direction which is specified under the null hypothesis. We are interested in the asymptotic behavior of some tests when both the sample size and the dimension become large.
