Distributionally robust L1-estimation in multiple linear regression

Linear regression is one of the most important and widely used techniques in data analysis, for which a key step is the estimation of the unknown parameters. However, it is often carried out under the assumption that the full information of the error distribution is available. This is clearly unreal...

Full description

Bibliographic Details
Main Authors: Gong, Z., Liu, C., Sun, Jie, Teo, Kok Lay
Format: Journal Article
Published: Springer Verlag 2018
Online Access:http://purl.org/au-research/grants/arc/DP160102819
http://hdl.handle.net/20.500.11937/73292
_version_ 1848762975716900864
author Gong, Z.
Liu, C.
Sun, Jie
Teo, Kok Lay
author_facet Gong, Z.
Liu, C.
Sun, Jie
Teo, Kok Lay
author_sort Gong, Z.
building Curtin Institutional Repository
collection Online Access
description Linear regression is one of the most important and widely used techniques in data analysis, for which a key step is the estimation of the unknown parameters. However, it is often carried out under the assumption that the full information of the error distribution is available. This is clearly unrealistic in practice. In this paper, we propose a distributionally robust formulation of L1-estimation (or the least absolute value estimation) problem, where the only knowledge on the error distribution is that it belongs to a well-defined ambiguity set. We then reformulate the estimation problem as a computationally tractable conic optimization problem by using duality theory. Finally, a numerical example is solved as a conic optimization problem to demonstrate the effectiveness of the proposed approach.
first_indexed 2025-11-14T10:56:07Z
format Journal Article
id curtin-20.500.11937-73292
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T10:56:07Z
publishDate 2018
publisher Springer Verlag
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-732922022-10-27T04:49:04Z Distributionally robust L1-estimation in multiple linear regression Gong, Z. Liu, C. Sun, Jie Teo, Kok Lay Linear regression is one of the most important and widely used techniques in data analysis, for which a key step is the estimation of the unknown parameters. However, it is often carried out under the assumption that the full information of the error distribution is available. This is clearly unrealistic in practice. In this paper, we propose a distributionally robust formulation of L1-estimation (or the least absolute value estimation) problem, where the only knowledge on the error distribution is that it belongs to a well-defined ambiguity set. We then reformulate the estimation problem as a computationally tractable conic optimization problem by using duality theory. Finally, a numerical example is solved as a conic optimization problem to demonstrate the effectiveness of the proposed approach. 2018 Journal Article http://hdl.handle.net/20.500.11937/73292 10.1007/s11590-018-1299-x http://purl.org/au-research/grants/arc/DP160102819 http://purl.org/au-research/grants/arc/DP140100289 Springer Verlag fulltext
spellingShingle Gong, Z.
Liu, C.
Sun, Jie
Teo, Kok Lay
Distributionally robust L1-estimation in multiple linear regression
title Distributionally robust L1-estimation in multiple linear regression
title_full Distributionally robust L1-estimation in multiple linear regression
title_fullStr Distributionally robust L1-estimation in multiple linear regression
title_full_unstemmed Distributionally robust L1-estimation in multiple linear regression
title_short Distributionally robust L1-estimation in multiple linear regression
title_sort distributionally robust l1-estimation in multiple linear regression
url http://purl.org/au-research/grants/arc/DP160102819
http://purl.org/au-research/grants/arc/DP160102819
http://hdl.handle.net/20.500.11937/73292