Robust Signal Recovery for High-Dimensional Linear Log-Contrast Models with Compositional Covariates

Dongxiao Han; Jian Huang; Yuanyuan Lin; Lei Liu; Lianqiang Qu; Liuquan Sun

doi:10.1080/07350015.2022.2097911

Back

Robust Signal Recovery for High-Dimensional Linear Log-Contrast Models with Compositional Covariates

Journal article

Peer reviewed

Robust Signal Recovery for High-Dimensional Linear Log-Contrast Models with Compositional Covariates

Dongxiao Han, Jian Huang, Yuanyuan Lin, Lei Liu, Lianqiang Qu and Liuquan Sun

Journal of business & economic statistics, Vol.ahead-of-print(ahead-of-print), pp.1-11

09/13/2022

DOI: 10.1080/07350015.2022.2097911

Files and links (1)

url

https://figshare.com/articles/journal_contribution/Robust_Signal_Recovery_for_High-dimensional_Linear_Log-contrast_Models_with_Compositional_Covariates/20238934View

Open Access

Abstract

In this article, we propose a robust signal recovery method for high-dimensional linear log-contrast models, when the error distribution could be heavy-tailed and asymmetric. The proposed method is built on the Huber loss with penalization. We establish the and consistency for the resulting estimator. Under conditions analogous to the irrepresentability condition and the minimum signal strength condition, we prove that the signed support of the slope parameter vector can be recovered with high probability. The finite-sample behavior of the proposed method is evaluated through simulation studies, and applications to a GDP satisfaction dataset an HIV microbiome dataset are provided.

Compositional data

Consistent estimation

Huber loss

Lasso

Support recovery

Details

Title: Subtitle: Robust Signal Recovery for High-Dimensional Linear Log-Contrast Models with Compositional Covariates
Creators: Dongxiao Han - Nankai University
Jian Huang - University of Iowa
Yuanyuan Lin - Chinese University of Hong Kong
Lei Liu - Washington University in St. Louis
Lianqiang Qu - Central China Normal University
Liuquan Sun - Academy of Mathematics and Systems Science
Resource Type: Journal article
Publication Details: Journal of business & economic statistics, Vol.ahead-of-print(ahead-of-print), pp.1-11
DOI: 10.1080/07350015.2022.2097911
ISSN: 0735-0015
eISSN: 1537-2707
Publisher: Taylor & Francis
Grant note: DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 12101330; name: Central Universities, Nankai University, award: 9920200110; DOI: 10.13039/100000001, name: U.S. National Science Foundation, award: DMS-1916199; name: Hong Kong Research Grants Council, award: 14306219, 14306620; DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 11961028; DOI: 10.13039/501100004853, name: Chinese University of Hong Kong., award: TR002345; DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 12001219; DOI: 10.13039/501100001809, name: National Natural Science Foundation of China, award: 12171463
Language: English
Electronic publication date: 09/13/2022
Academic Unit: Statistics and Actuarial Science
Record Identifier: 9984296160802771

Metrics

94 Record Views

4 Times Cited - Web of Science