Understanding covariate shift in model performance

Three (3) different methods (logistic regression, covariate shift and k-NN) were applied to five (5) internal datasets and one (1) external, publically available dataset where covariate shift existed. In all cases, k-NN’s performance was inferior to either logistic regression or covariate shift. Sur...

Descripción completa

Detalles Bibliográficos
Autores principales: McGaughey, Georgia, Walters, W. Patrick, Goldman, Brian
Formato: Online Artículo Texto
Lenguaje:English
Publicado: F1000Research 2016
Materias:
Research Note
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5070592/
https://www.ncbi.nlm.nih.gov/pubmed/27803797
http://dx.doi.org/10.12688/f1000research.8317.3
_version_ 1782461168144613376
author McGaughey, Georgia
Walters, W. Patrick
Goldman, Brian
author_facet McGaughey, Georgia
Walters, W. Patrick
Goldman, Brian
author_sort McGaughey, Georgia
collection PubMed
description Three (3) different methods (logistic regression, covariate shift and k-NN) were applied to five (5) internal datasets and one (1) external, publically available dataset where covariate shift existed. In all cases, k-NN’s performance was inferior to either logistic regression or covariate shift. Surprisingly, there was no obvious advantage for using covariate shift to reweight the training data in the examined datasets.
format Online
Article
Text
id pubmed-5070592
institution National Center for Biotechnology Information
language English
publishDate 2016
publisher F1000Research
record_format MEDLINE/PubMed
spelling pubmed-50705922016-10-31 Understanding covariate shift in model performance McGaughey, Georgia Walters, W. Patrick Goldman, Brian F1000Res Research Note Three (3) different methods (logistic regression, covariate shift and k-NN) were applied to five (5) internal datasets and one (1) external, publically available dataset where covariate shift existed. In all cases, k-NN’s performance was inferior to either logistic regression or covariate shift. Surprisingly, there was no obvious advantage for using covariate shift to reweight the training data in the examined datasets. F1000Research 2016-10-17 /pmc/articles/PMC5070592/ /pubmed/27803797 http://dx.doi.org/10.12688/f1000research.8317.3 Text en Copyright: © 2016 McGaughey G et al. http://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
spellingShingle Research Note
McGaughey, Georgia
Walters, W. Patrick
Goldman, Brian
Understanding covariate shift in model performance
title Understanding covariate shift in model performance
title_full Understanding covariate shift in model performance
title_fullStr Understanding covariate shift in model performance
title_full_unstemmed Understanding covariate shift in model performance
title_short Understanding covariate shift in model performance
title_sort understanding covariate shift in model performance
topic Research Note
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5070592/
https://www.ncbi.nlm.nih.gov/pubmed/27803797
http://dx.doi.org/10.12688/f1000research.8317.3
work_keys_str_mv AT mcgaugheygeorgia understandingcovariateshiftinmodelperformance
AT walterswpatrick understandingcovariateshiftinmodelperformance
AT goldmanbrian understandingcovariateshiftinmodelperformance

Ejemplares similares