Error estimation and adaptive tuning for unregularized robust M-estimator
We consider unregularized robust M-estimators for linear models under Gaussian design and heavy-tailed noise, in the proportional asymptotics regime where the sample size n and the number of features p are both increasing such that $p/n \to \gamma\in (0,1)$. An estimator of the out-of-sample error o...
Gespeichert in:
Hauptverfasser: | , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | |
container_volume | |
creator | Bellec, Pierre C Koriyama, Takuya |
description | We consider unregularized robust M-estimators for linear models under
Gaussian design and heavy-tailed noise, in the proportional asymptotics regime
where the sample size n and the number of features p are both increasing such
that $p/n \to \gamma\in (0,1)$. An estimator of the out-of-sample error of a
robust M-estimator is analysed and proved to be consistent for a large family
of loss functions that includes the Huber loss. As an application of this
result, we propose an adaptive tuning procedure of the scale parameter
$\lambda>0$ of a given loss function $\rho$: choosing$\hat \lambda$ in a given
interval $I$ that minimizes the out-of-sample error estimate of the M-estimator
constructed with loss $\rho_\lambda(\cdot) = \lambda^2 \rho(\cdot/\lambda)$
leads to the optimal out-of-sample error over $I$. The proof relies on a
smoothing argument: the unregularized M-estimation objective function is
perturbed, or smoothed, with a Ridge penalty that vanishes as $n\to+\infty$,
and show that the unregularized M-estimator of interest inherits properties of
its smoothed version. |
doi_str_mv | 10.48550/arxiv.2312.13257 |
format | Article |
fullrecord | <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2312_13257</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2312_13257</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-f58d79d5e5a09ce4a2dfb41661c683355cae904fdc532b4611b26ff6339429383</originalsourceid><addsrcrecordid>eNotj71uwjAUhb0wVMADdKpfIGnsazvxWCFakKi6sEc3sY0sUQfdOKjt0_NTprOc8x19jD2LqlSN1tUr0k88lxKELAVIXT-x7ZpoIO7HHL8xxyFxTI6jw1OOZ8_zlGI68HCtTIn8YToixT_vOA3dNGb-WTyWAy3YLOBx9MtHztn-fb1fbYrd18d29bYr0NR1EXTjauu011jZ3iuULnRKGCN60wBo3aO3lQqu1yA7ZYTopAnBAFglLTQwZy__2LtLe6LrO_22N6f27gQXiAVHjQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Error estimation and adaptive tuning for unregularized robust M-estimator</title><source>arXiv.org</source><creator>Bellec, Pierre C ; Koriyama, Takuya</creator><creatorcontrib>Bellec, Pierre C ; Koriyama, Takuya</creatorcontrib><description>We consider unregularized robust M-estimators for linear models under
Gaussian design and heavy-tailed noise, in the proportional asymptotics regime
where the sample size n and the number of features p are both increasing such
that $p/n \to \gamma\in (0,1)$. An estimator of the out-of-sample error of a
robust M-estimator is analysed and proved to be consistent for a large family
of loss functions that includes the Huber loss. As an application of this
result, we propose an adaptive tuning procedure of the scale parameter
$\lambda>0$ of a given loss function $\rho$: choosing$\hat \lambda$ in a given
interval $I$ that minimizes the out-of-sample error estimate of the M-estimator
constructed with loss $\rho_\lambda(\cdot) = \lambda^2 \rho(\cdot/\lambda)$
leads to the optimal out-of-sample error over $I$. The proof relies on a
smoothing argument: the unregularized M-estimation objective function is
perturbed, or smoothed, with a Ridge penalty that vanishes as $n\to+\infty$,
and show that the unregularized M-estimator of interest inherits properties of
its smoothed version.</description><identifier>DOI: 10.48550/arxiv.2312.13257</identifier><language>eng</language><subject>Mathematics - Statistics Theory ; Statistics - Theory</subject><creationdate>2023-12</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2312.13257$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2312.13257$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Bellec, Pierre C</creatorcontrib><creatorcontrib>Koriyama, Takuya</creatorcontrib><title>Error estimation and adaptive tuning for unregularized robust M-estimator</title><description>We consider unregularized robust M-estimators for linear models under
Gaussian design and heavy-tailed noise, in the proportional asymptotics regime
where the sample size n and the number of features p are both increasing such
that $p/n \to \gamma\in (0,1)$. An estimator of the out-of-sample error of a
robust M-estimator is analysed and proved to be consistent for a large family
of loss functions that includes the Huber loss. As an application of this
result, we propose an adaptive tuning procedure of the scale parameter
$\lambda>0$ of a given loss function $\rho$: choosing$\hat \lambda$ in a given
interval $I$ that minimizes the out-of-sample error estimate of the M-estimator
constructed with loss $\rho_\lambda(\cdot) = \lambda^2 \rho(\cdot/\lambda)$
leads to the optimal out-of-sample error over $I$. The proof relies on a
smoothing argument: the unregularized M-estimation objective function is
perturbed, or smoothed, with a Ridge penalty that vanishes as $n\to+\infty$,
and show that the unregularized M-estimator of interest inherits properties of
its smoothed version.</description><subject>Mathematics - Statistics Theory</subject><subject>Statistics - Theory</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj71uwjAUhb0wVMADdKpfIGnsazvxWCFakKi6sEc3sY0sUQfdOKjt0_NTprOc8x19jD2LqlSN1tUr0k88lxKELAVIXT-x7ZpoIO7HHL8xxyFxTI6jw1OOZ8_zlGI68HCtTIn8YToixT_vOA3dNGb-WTyWAy3YLOBx9MtHztn-fb1fbYrd18d29bYr0NR1EXTjauu011jZ3iuULnRKGCN60wBo3aO3lQqu1yA7ZYTopAnBAFglLTQwZy__2LtLe6LrO_22N6f27gQXiAVHjQ</recordid><startdate>20231220</startdate><enddate>20231220</enddate><creator>Bellec, Pierre C</creator><creator>Koriyama, Takuya</creator><scope>AKZ</scope><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20231220</creationdate><title>Error estimation and adaptive tuning for unregularized robust M-estimator</title><author>Bellec, Pierre C ; Koriyama, Takuya</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-f58d79d5e5a09ce4a2dfb41661c683355cae904fdc532b4611b26ff6339429383</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Mathematics - Statistics Theory</topic><topic>Statistics - Theory</topic><toplevel>online_resources</toplevel><creatorcontrib>Bellec, Pierre C</creatorcontrib><creatorcontrib>Koriyama, Takuya</creatorcontrib><collection>arXiv Mathematics</collection><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Bellec, Pierre C</au><au>Koriyama, Takuya</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Error estimation and adaptive tuning for unregularized robust M-estimator</atitle><date>2023-12-20</date><risdate>2023</risdate><abstract>We consider unregularized robust M-estimators for linear models under
Gaussian design and heavy-tailed noise, in the proportional asymptotics regime
where the sample size n and the number of features p are both increasing such
that $p/n \to \gamma\in (0,1)$. An estimator of the out-of-sample error of a
robust M-estimator is analysed and proved to be consistent for a large family
of loss functions that includes the Huber loss. As an application of this
result, we propose an adaptive tuning procedure of the scale parameter
$\lambda>0$ of a given loss function $\rho$: choosing$\hat \lambda$ in a given
interval $I$ that minimizes the out-of-sample error estimate of the M-estimator
constructed with loss $\rho_\lambda(\cdot) = \lambda^2 \rho(\cdot/\lambda)$
leads to the optimal out-of-sample error over $I$. The proof relies on a
smoothing argument: the unregularized M-estimation objective function is
perturbed, or smoothed, with a Ridge penalty that vanishes as $n\to+\infty$,
and show that the unregularized M-estimator of interest inherits properties of
its smoothed version.</abstract><doi>10.48550/arxiv.2312.13257</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | DOI: 10.48550/arxiv.2312.13257 |
ispartof | |
issn | |
language | eng |
recordid | cdi_arxiv_primary_2312_13257 |
source | arXiv.org |
subjects | Mathematics - Statistics Theory Statistics - Theory |
title | Error estimation and adaptive tuning for unregularized robust M-estimator |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-26T11%3A50%3A56IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Error%20estimation%20and%20adaptive%20tuning%20for%20unregularized%20robust%20M-estimator&rft.au=Bellec,%20Pierre%20C&rft.date=2023-12-20&rft_id=info:doi/10.48550/arxiv.2312.13257&rft_dat=%3Carxiv_GOX%3E2312_13257%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |