Tuning free Catoni type joint robust estimation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Xiang, Liu, Jun S., Sun, Qiang, Xu, Lihu
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917501512712192
author Li, Xiang
Liu, Jun S.
Sun, Qiang
Xu, Lihu
author_facet Li, Xiang
Liu, Jun S.
Sun, Qiang
Xu, Lihu
contents This paper develops a Catoni-type joint (tuning-free) estimation framework for parametric models with heavy-tailed noise, in which the target parameter and the unknown noise variance are estimated simultaneously through a system of two coupled Catoni-type estimating equations. We instantiate the framework in three canonical settings: mean estimation, linear regression, and $\ell_{2}$-penalized regression. Theoretically, we establish non-asymptotic, sub-Gaussian-type deviation bounds that hold jointly for the target parameter and the variance estimator, under only a finite $2β$-th moment assumption with $β\in (1,2]$. The resulting rates match -- up to absolute constants -- those of oracle procedures that know the variance in advance, thereby attaining optimality in the heavy-tailed regime. Methodologically, because the coupled equations are intrinsically non-convex and non-linear, classical convex M-estimation arguments are inapplicable. We develop a new analytical toolkit based on the Poincare--Miranda theorem. The resulting proof strategy is of independent methodological interest, and we expect it to be applicable to a broad class of other statistical problems in which several parameters of heterogeneous nature must be estimated jointly.
format Preprint
id arxiv_https___arxiv_org_abs_2511_11054
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Tuning free Catoni type joint robust estimation
Li, Xiang
Liu, Jun S.
Sun, Qiang
Xu, Lihu
Statistics Theory
Primary 62G35, 62F35, Secondary 62J07, 62H12
This paper develops a Catoni-type joint (tuning-free) estimation framework for parametric models with heavy-tailed noise, in which the target parameter and the unknown noise variance are estimated simultaneously through a system of two coupled Catoni-type estimating equations. We instantiate the framework in three canonical settings: mean estimation, linear regression, and $\ell_{2}$-penalized regression. Theoretically, we establish non-asymptotic, sub-Gaussian-type deviation bounds that hold jointly for the target parameter and the variance estimator, under only a finite $2β$-th moment assumption with $β\in (1,2]$. The resulting rates match -- up to absolute constants -- those of oracle procedures that know the variance in advance, thereby attaining optimality in the heavy-tailed regime. Methodologically, because the coupled equations are intrinsically non-convex and non-linear, classical convex M-estimation arguments are inapplicable. We develop a new analytical toolkit based on the Poincare--Miranda theorem. The resulting proof strategy is of independent methodological interest, and we expect it to be applicable to a broad class of other statistical problems in which several parameters of heterogeneous nature must be estimated jointly.
title Tuning free Catoni type joint robust estimation
topic Statistics Theory
Primary 62G35, 62F35, Secondary 62J07, 62H12
url https://arxiv.org/abs/2511.11054