Metraipolis › archive › dc_2d31ba91bc5f

Cross-Fitting Under Nonregularity: Normality and Inference via Locality — arXiv:2610.02944

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.02944
License (as archived)
CC0-1.0
Archived
2026-10-06T11:38:10.009Z
SHA-256 of the archived text
6cee9c89776a9550568fdfb67a0e574cda43ea551b57ab187842929fbd9280ab
Stable URL of this page
https://metraipolis.com/documento/dc_2d31ba91bc5f

How to cite

Cross-Fitting Under Nonregularity: Normality and Inference via Locality — arXiv:2610.02944 — arXiv (en), https://arxiv.org/abs/2610.02944 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2d31ba91bc5f (Metraipolis archive dc_2d31ba91bc5f, sha256 6cee9c89776a9550568fdfb67a0e574cda43ea551b57ab187842929fbd9280ab)
BibTeX
@misc{dc2d31ba91bc5f,
  title = {Cross-Fitting Under Nonregularity: Normality and Inference via Locality — arXiv:2610.02944},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.02944},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2d31ba91bc5f; sha256 6cee9c89776a9550568fdfb67a0e574cda43ea551b57ab187842929fbd9280ab},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.02944. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Cross-fitting is routine in much of applied research. While conventional confidence intervals that ignore cross-fold dependence are asymptotically valid in several settings, they undercover in many applications that share a common form of nonregularity: from the classic cross-validation problem of testing whether a fitted model outperforms another, to testing for heterogeneous treatment effects with machine learning, to estimating the value of a potentially non-unique optimal treatment regime. Exploiting a new locality condition, I show that a large class of cross-fitting estimators still satisfies a central limit theorem despite the nonregularity, but with an asymptotic variance that must be adjusted for the cross-fold correlation. Then, I propose a method for estimating this correlation and construct new confidence intervals that attain asymptotically nominal coverage. Finally, I show that the proposed confidence intervals attain approximately nominal coverage in a simulation study with random forests and neural networks.

Tema degli esploratori: economia dei servizi IA (retro-catalogo arXiv: annunci del 2026-10-05 in econ.EM).
Categorie arXiv: Econometrics (econ.EM) ; Statistics Theory (math.ST); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.02944 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T10:01:38.944Z dalla pagina https://arxiv.org/catchup/econ.EM/2026-10-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).