Design of c-optimal experiments for high-dimensional linear models

Hamid Eftekhari; Moulinath Banerjee; Ya’Acov Ritov

doi:10.3150/22-BEJ1472

Design of c-optimal experiments for high-dimensional linear models

Hamid Eftekhari, Moulinath Banerjee, Ya’Acov Ritov

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

We study randomized designs that minimize the asymptotic variance of a debiased lasso estimator when a large pool of unlabeled data is available but measuring the corresponding responses is costly. The optimal sampling distribution arises as the solution of a semidefinite program. The improvements in efficiency that result from these optimal designs are demonstrated via simulation experiments.

Original language	English
Pages (from-to)	652-668
Number of pages	17
Journal	Bernoulli
Volume	29
Issue number	1
DOIs	https://doi.org/10.3150/22-BEJ1472
State	Published - Feb 2023
Externally published	Yes

Bibliographical note

Publisher Copyright:
© 2023 ISI/BS.

Keywords

compressed sensing
inference
Optimal design
sparsity

Access to Document

10.3150/22-BEJ1472

Cite this

@article{9fc00140e8dc42e7bf81e4c022bcfaf6,

title = "Design of c-optimal experiments for high-dimensional linear models",

abstract = "We study randomized designs that minimize the asymptotic variance of a debiased lasso estimator when a large pool of unlabeled data is available but measuring the corresponding responses is costly. The optimal sampling distribution arises as the solution of a semidefinite program. The improvements in efficiency that result from these optimal designs are demonstrated via simulation experiments.",

keywords = "compressed sensing, inference, Optimal design, sparsity",

author = "Hamid Eftekhari and Moulinath Banerjee and Ya{\textquoteright}Acov Ritov",

note = "Publisher Copyright: {\textcopyright} 2023 ISI/BS.",

year = "2023",

month = feb,

doi = "10.3150/22-BEJ1472",

language = "אנגלית",

volume = "29",

pages = "652--668",

journal = "Bernoulli",

issn = "1350-7265",

publisher = "Bernoulli Society for Mathematical Statistics and Probability",

number = "1",

}

TY - JOUR

T1 - Design of c-optimal experiments for high-dimensional linear models

AU - Eftekhari, Hamid

AU - Banerjee, Moulinath

AU - Ritov, Ya’Acov

PY - 2023/2

Y1 - 2023/2

N2 - We study randomized designs that minimize the asymptotic variance of a debiased lasso estimator when a large pool of unlabeled data is available but measuring the corresponding responses is costly. The optimal sampling distribution arises as the solution of a semidefinite program. The improvements in efficiency that result from these optimal designs are demonstrated via simulation experiments.

AB - We study randomized designs that minimize the asymptotic variance of a debiased lasso estimator when a large pool of unlabeled data is available but measuring the corresponding responses is costly. The optimal sampling distribution arises as the solution of a semidefinite program. The improvements in efficiency that result from these optimal designs are demonstrated via simulation experiments.

KW - compressed sensing

KW - inference

KW - Optimal design

KW - sparsity

UR - http://www.scopus.com/inward/record.url?scp=85139971661&partnerID=8YFLogxK

U2 - 10.3150/22-BEJ1472

DO - 10.3150/22-BEJ1472

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:85139971661

SN - 1350-7265

VL - 29

SP - 652

EP - 668

JO - Bernoulli

JF - Bernoulli

IS - 1

ER -

Design of c-optimal experiments for high-dimensional linear models

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this