A carregar...

Fold-stratified cross-validation for unbiased and privacy-preserving federated learning

OBJECTIVE: We introduce fold-stratified cross-validation, a validation methodology that is compatible with privacy-preserving federated learning and that prevents data leakage caused by duplicates of electronic health records (EHRs). MATERIALS AND METHODS: Fold-stratified cross-validation complement...

ver descrição completa

Na minha lista:
Detalhes bibliográficos
Publicado no:J Am Med Inform Assoc
Main Authors: Bey, Romain, Goussault, Romain, Grolleau, François, Benchoufi, Mehdi, Porcher, Raphaël
Formato: Artigo
Idioma:Inglês
Publicado em: Oxford University Press 2020
Assuntos:
Acesso em linha:https://ncbi.nlm.nih.gov/pmc/articles/PMC7647321/
https://ncbi.nlm.nih.gov/pubmed/32620945
https://ncbi.nlm.nih.govhttp://dx.doi.org/10.1093/jamia/ocaa096
Tags: Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!