English

Wald Statistics in high-dimensional PCA

Statistics Theory 2018-05-11 v1 Statistics Theory

Abstract

In this note we consider PCA for Gaussian observations X1,,XnX_1,\dots, X_n with covariance Σ=iλiPi\Sigma=\sum_i \lambda_i P_i in the 'effective rank' setting with model complexity governed by r(Σ):=tr(Σ)/Σ\mathbf{r}(\Sigma):=\text{tr}(\Sigma)/\| \Sigma \|. We prove a Berry-Essen type bound for a Wald Statistic of the spectral projector P^r\hat P_r. This can be used to construct non-asymptotic confidence ellipsoids and tests for spectral projectors PrP_r. Using higher order pertubation theory we are able to show that our Theorem remains valid even when r(Σ)n\mathbf{r}(\Sigma) \gg \sqrt{n}.

Keywords

Cite

@article{arxiv.1805.03839,
  title  = {Wald Statistics in high-dimensional PCA},
  author = {Matthias Löffler},
  journal= {arXiv preprint arXiv:1805.03839},
  year   = {2018}
}

Comments

11 pages

R2 v1 2026-06-23T01:50:38.524Z