Cargando…
Regression on imperfect class labels derived by unsupervised clustering
Outcome regressed on class labels identified by unsupervised clustering is custom in many applications. However, it is common to ignore the misclassification of class labels caused by the learning algorithm, which potentially leads to serious bias of the estimated effect parameters. Due to their gen...
Autores principales: | , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Oxford University Press
2020
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7986660/ https://www.ncbi.nlm.nih.gov/pubmed/32124917 http://dx.doi.org/10.1093/bib/bbaa014 |
_version_ | 1783668485655625728 |
---|---|
author | Brøndum, Rasmus Froberg Michaelsen, Thomas Yssing Bøgsted, Martin |
author_facet | Brøndum, Rasmus Froberg Michaelsen, Thomas Yssing Bøgsted, Martin |
author_sort | Brøndum, Rasmus Froberg |
collection | PubMed |
description | Outcome regressed on class labels identified by unsupervised clustering is custom in many applications. However, it is common to ignore the misclassification of class labels caused by the learning algorithm, which potentially leads to serious bias of the estimated effect parameters. Due to their generality we suggest to address the problem by use of regression calibration or the misclassification simulation and extrapolation method. Performance is illustrated by simulated data from Gaussian mixture models, documenting a reduced bias and improved coverage of confidence intervals when adjusting for misclassification with either method. Finally, we apply our method to data from a previous study, which regressed overall survival on class labels derived from unsupervised clustering of gene expression data from bone marrow samples of multiple myeloma patients. |
format | Online Article Text |
id | pubmed-7986660 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2020 |
publisher | Oxford University Press |
record_format | MEDLINE/PubMed |
spelling | pubmed-79866602021-03-26 Regression on imperfect class labels derived by unsupervised clustering Brøndum, Rasmus Froberg Michaelsen, Thomas Yssing Bøgsted, Martin Brief Bioinform Problem Solving Protocol Outcome regressed on class labels identified by unsupervised clustering is custom in many applications. However, it is common to ignore the misclassification of class labels caused by the learning algorithm, which potentially leads to serious bias of the estimated effect parameters. Due to their generality we suggest to address the problem by use of regression calibration or the misclassification simulation and extrapolation method. Performance is illustrated by simulated data from Gaussian mixture models, documenting a reduced bias and improved coverage of confidence intervals when adjusting for misclassification with either method. Finally, we apply our method to data from a previous study, which regressed overall survival on class labels derived from unsupervised clustering of gene expression data from bone marrow samples of multiple myeloma patients. Oxford University Press 2020-03-03 /pmc/articles/PMC7986660/ /pubmed/32124917 http://dx.doi.org/10.1093/bib/bbaa014 Text en © The Author(s) 2020. Published by Oxford University Press. https://creativecommons.org/licenses/by/4.0/This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) ), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. |
spellingShingle | Problem Solving Protocol Brøndum, Rasmus Froberg Michaelsen, Thomas Yssing Bøgsted, Martin Regression on imperfect class labels derived by unsupervised clustering |
title | Regression on imperfect class labels derived by unsupervised clustering |
title_full | Regression on imperfect class labels derived by unsupervised clustering |
title_fullStr | Regression on imperfect class labels derived by unsupervised clustering |
title_full_unstemmed | Regression on imperfect class labels derived by unsupervised clustering |
title_short | Regression on imperfect class labels derived by unsupervised clustering |
title_sort | regression on imperfect class labels derived by unsupervised clustering |
topic | Problem Solving Protocol |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7986660/ https://www.ncbi.nlm.nih.gov/pubmed/32124917 http://dx.doi.org/10.1093/bib/bbaa014 |
work_keys_str_mv | AT brøndumrasmusfroberg regressiononimperfectclasslabelsderivedbyunsupervisedclustering AT michaelsenthomasyssing regressiononimperfectclasslabelsderivedbyunsupervisedclustering AT bøgstedmartin regressiononimperfectclasslabelsderivedbyunsupervisedclustering |