Cargando…

Cooperative learning for multiview analysis

We propose a method for supervised learning with multiple sets of features (“views”). The multiview problem is especially important in biology and medicine, where “-omics” data, such as genomics, proteomics, and radiomics, are measured on a common set of samples. “Cooperative learning” combines the...

Descripción completa

Detalles Bibliográficos
Autores principales: Ding, Daisy Yi, Li, Shuangning, Narasimhan, Balasubramanian, Tibshirani, Robert
Formato: Online Artículo Texto
Lenguaje:English
Publicado: National Academy of Sciences 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9499553/
https://www.ncbi.nlm.nih.gov/pubmed/36095183
http://dx.doi.org/10.1073/pnas.2202113119
_version_ 1784795019589189632
author Ding, Daisy Yi
Li, Shuangning
Narasimhan, Balasubramanian
Tibshirani, Robert
author_facet Ding, Daisy Yi
Li, Shuangning
Narasimhan, Balasubramanian
Tibshirani, Robert
author_sort Ding, Daisy Yi
collection PubMed
description We propose a method for supervised learning with multiple sets of features (“views”). The multiview problem is especially important in biology and medicine, where “-omics” data, such as genomics, proteomics, and radiomics, are measured on a common set of samples. “Cooperative learning” combines the usual squared-error loss of predictions with an “agreement” penalty to encourage the predictions from different data views to agree. By varying the weight of the agreement penalty, we get a continuum of solutions that include the well-known early and late fusion approaches. Cooperative learning chooses the degree of agreement (or fusion) in an adaptive manner, using a validation set or cross-validation to estimate test set prediction error. One version of our fitting procedure is modular, where one can choose different fitting mechanisms (e.g., lasso, random forests, boosting, or neural networks) appropriate for different data views. In the setting of cooperative regularized linear regression, the method combines the lasso penalty with the agreement penalty, yielding feature sparsity. The method can be especially powerful when the different data views share some underlying relationship in their signals that can be exploited to boost the signals. We show that cooperative learning achieves higher predictive accuracy on simulated data and real multiomics examples of labor-onset prediction. By leveraging aligned signals and allowing flexible fitting mechanisms for different modalities, cooperative learning offers a powerful approach to multiomics data fusion.
format Online
Article
Text
id pubmed-9499553
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher National Academy of Sciences
record_format MEDLINE/PubMed
spelling pubmed-94995532023-03-12 Cooperative learning for multiview analysis Ding, Daisy Yi Li, Shuangning Narasimhan, Balasubramanian Tibshirani, Robert Proc Natl Acad Sci U S A Physical Sciences We propose a method for supervised learning with multiple sets of features (“views”). The multiview problem is especially important in biology and medicine, where “-omics” data, such as genomics, proteomics, and radiomics, are measured on a common set of samples. “Cooperative learning” combines the usual squared-error loss of predictions with an “agreement” penalty to encourage the predictions from different data views to agree. By varying the weight of the agreement penalty, we get a continuum of solutions that include the well-known early and late fusion approaches. Cooperative learning chooses the degree of agreement (or fusion) in an adaptive manner, using a validation set or cross-validation to estimate test set prediction error. One version of our fitting procedure is modular, where one can choose different fitting mechanisms (e.g., lasso, random forests, boosting, or neural networks) appropriate for different data views. In the setting of cooperative regularized linear regression, the method combines the lasso penalty with the agreement penalty, yielding feature sparsity. The method can be especially powerful when the different data views share some underlying relationship in their signals that can be exploited to boost the signals. We show that cooperative learning achieves higher predictive accuracy on simulated data and real multiomics examples of labor-onset prediction. By leveraging aligned signals and allowing flexible fitting mechanisms for different modalities, cooperative learning offers a powerful approach to multiomics data fusion. National Academy of Sciences 2022-09-12 2022-09-20 /pmc/articles/PMC9499553/ /pubmed/36095183 http://dx.doi.org/10.1073/pnas.2202113119 Text en Copyright © 2022 the Author(s). Published by PNAS. https://creativecommons.org/licenses/by-nc-nd/4.0/This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND) (https://creativecommons.org/licenses/by-nc-nd/4.0/) .
spellingShingle Physical Sciences
Ding, Daisy Yi
Li, Shuangning
Narasimhan, Balasubramanian
Tibshirani, Robert
Cooperative learning for multiview analysis
title Cooperative learning for multiview analysis
title_full Cooperative learning for multiview analysis
title_fullStr Cooperative learning for multiview analysis
title_full_unstemmed Cooperative learning for multiview analysis
title_short Cooperative learning for multiview analysis
title_sort cooperative learning for multiview analysis
topic Physical Sciences
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9499553/
https://www.ncbi.nlm.nih.gov/pubmed/36095183
http://dx.doi.org/10.1073/pnas.2202113119
work_keys_str_mv AT dingdaisyyi cooperativelearningformultiviewanalysis
AT lishuangning cooperativelearningformultiviewanalysis
AT narasimhanbalasubramanian cooperativelearningformultiviewanalysis
AT tibshiranirobert cooperativelearningformultiviewanalysis