Topics in robust statistical learning*
Some recent contributions to robust inference are presented. Firstly, the classical problem of robust M-estimation of a location parameter is revisited using an optimal transport approach - with specifically designed Wasserstein-type distances - that reduces robustness to a continuity property. Seco...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
EDP Sciences
2023-11-01
|
Series: | ESAIM: Proceedings and Surveys |
Online Access: | https://www.esaim-proc.org/articles/proc/pdf/2023/03/proc230808.pdf |
_version_ | 1797343807824461824 |
---|---|
author | Brecheteau Claire Genetay Edouard Mathieu Timothee Saumard Adrien |
author_facet | Brecheteau Claire Genetay Edouard Mathieu Timothee Saumard Adrien |
author_sort | Brecheteau Claire |
collection | DOAJ |
description | Some recent contributions to robust inference are presented. Firstly, the classical problem of robust M-estimation of a location parameter is revisited using an optimal transport approach - with specifically designed Wasserstein-type distances - that reduces robustness to a continuity property. Secondly, a procedure of estimation of the distance function to a compact set is described, using union of balls. This methodology originates in the field of topological inference and offers as a byproduct a robust clustering method. Thirdly, a robust Lloyd-type algorithm for clustering is constructed, using a bootstrap variant of the median-of-means strategy. This algorithm comes with a robust initialization. |
first_indexed | 2024-03-08T10:53:09Z |
format | Article |
id | doaj.art-31095b126eca43378524c3facaeae5e0 |
institution | Directory Open Access Journal |
issn | 2267-3059 |
language | English |
last_indexed | 2024-03-08T10:53:09Z |
publishDate | 2023-11-01 |
publisher | EDP Sciences |
record_format | Article |
series | ESAIM: Proceedings and Surveys |
spelling | doaj.art-31095b126eca43378524c3facaeae5e02024-01-26T16:41:54ZengEDP SciencesESAIM: Proceedings and Surveys2267-30592023-11-017411913610.1051/proc/202374119proc230808Topics in robust statistical learning*Brecheteau Claire0Genetay Edouard1Mathieu Timothee2Saumard Adrien3Univ. Rennes 2CREST, ENSAI, Univ. Rennes, LumenAIINRIA, Scool team. Univ. Lille, CRIStAL, CNRSCREST, ENSAI, Univ. RennesSome recent contributions to robust inference are presented. Firstly, the classical problem of robust M-estimation of a location parameter is revisited using an optimal transport approach - with specifically designed Wasserstein-type distances - that reduces robustness to a continuity property. Secondly, a procedure of estimation of the distance function to a compact set is described, using union of balls. This methodology originates in the field of topological inference and offers as a byproduct a robust clustering method. Thirdly, a robust Lloyd-type algorithm for clustering is constructed, using a bootstrap variant of the median-of-means strategy. This algorithm comes with a robust initialization.https://www.esaim-proc.org/articles/proc/pdf/2023/03/proc230808.pdf |
spellingShingle | Brecheteau Claire Genetay Edouard Mathieu Timothee Saumard Adrien Topics in robust statistical learning* ESAIM: Proceedings and Surveys |
title | Topics in robust statistical learning* |
title_full | Topics in robust statistical learning* |
title_fullStr | Topics in robust statistical learning* |
title_full_unstemmed | Topics in robust statistical learning* |
title_short | Topics in robust statistical learning* |
title_sort | topics in robust statistical learning |
url | https://www.esaim-proc.org/articles/proc/pdf/2023/03/proc230808.pdf |
work_keys_str_mv | AT brecheteauclaire topicsinrobuststatisticallearning AT genetayedouard topicsinrobuststatisticallearning AT mathieutimothee topicsinrobuststatisticallearning AT saumardadrien topicsinrobuststatisticallearning |