English

Computable Bernstein Certificates for Cross-Fitted Clipped Covariance Estimation

Machine Learning 2026-02-17 v1 Machine Learning

Abstract

We study operator-norm covariance estimation from heavy-tailed samples that may include a small fraction of arbitrary outliers. A simple and widely used safeguard is \emph{Euclidean norm clipping}, but its accuracy depends critically on an unknown clipping level. We propose a cross-fitted clipped covariance estimator equipped with \emph{fully computable} Bernstein-type deviation certificates, enabling principled data-driven tuning via a selector (\emph{MinUpper}) that balances certified stochastic error and a robust hold-out proxy for clipping bias. The resulting procedure adapts to intrinsic complexity measures such as effective rank under mild tail regularity and retains meaningful guarantees under only finite fourth moments. Experiments on contaminated spiked-covariance benchmarks illustrate stable performance and competitive accuracy across regimes.

Keywords

Cite

@article{arxiv.2602.14020,
  title  = {Computable Bernstein Certificates for Cross-Fitted Clipped Covariance Estimation},
  author = {Even He and Zaizai Yan},
  journal= {arXiv preprint arXiv:2602.14020},
  year   = {2026}
}
R2 v1 2026-07-01T10:37:20.038Z