arXiv Analytics

Sign in

arXiv:1905.06318 [math.ST]AbstractReferencesReviewsResources

Which principal components are most sensitive to distributional changes?

Martin Tveten

Published 2019-05-15Version 1

PCA is often used in anomaly detection and statistical process control tasks. For bivariate data, we prove that the minor projection (the least varying projection) of the PCA-rotated data is the most sensitive to distributional changes, where sensitivity is defined by the Hellinger distance between distributions before and after a change. In particular, this is almost always the case if only one parameter of the bivariate normal distribution changes, i.e., the change is sparse. Simulations indicate that the minor projections are the most sensitive for a large range of changes and pre-change settings in higher dimensions as well. This motivates using the minor projections for detecting sparse distributional changes in high-dimensional data.

Related articles:
arXiv:1211.2117 [math.ST] (Published 2012-11-09)
Optimal rank-based testing for principal components
arXiv:0805.0465 [math.ST] (Published 2008-05-05)
Consistency of restricted maximum likelihood estimators of principal components
arXiv:1502.07042 [math.ST] (Published 2015-02-25)
Robust estimation of principal components from depth-based multivariate rank covariance matrix