Beyond the AUROC: data leakage, random splitting, and neglected calibration in critical care machine learning
Eurasian Journal of Critical Care · 10.08.2026 · 8 (2) · 113-114
DOI: 10.55994/ejcc.1993408 ↗Kurz gefasst
Fachartikel in der Zeitschrift Eurasian Journal of Critical Care, online erschienen am 10.08.2026. Verfasst von Ömer Faruk Çakıroğlu.
Abstract (Originalsprache)
Machine learning is increasingly proposed for predicting triage acuity, deterioration, and mortality in emergency and critical care, yet reported performance is often optimistic. Systematic reviews show that such models are frequently poorly reported and at high risk of bias. We highlight four recurring problems: data leakage, when preprocessing, feature selection, resampling, or hyperparameter tuning extends beyond the training partition; random splitting of a single dataset, which is neither external validation nor statistically efficient; evaluation confined to discrimination or accuracy, disregarding calibration and clinical utility; and unfair comparison of complex algorithms with simpler models and established clinical scores. We urge editors and reviewers to require adherence to current reporting and risk of bias standards, and to constrain claims of clinical readiness without external and prospective validation.
Zitieren
APA
Çakıroğlu, Ö. F. (2026). Beyond the AUROC: data leakage, random splitting, and neglected calibration in critical care machine learning. Eurasian Journal of Critical Care, 8(2), 113-114. https://doi.org/10.55994/ejcc.1993408
Vancouver
Çakıroğlu ÖF. Beyond the AUROC: data leakage, random splitting, and neglected calibration in critical care machine learning. Eurasian Journal of Critical Care. 2026;8(2):113-114. doi:10.55994/ejcc.1993408
BibTeX
@article{akrolu2026,
author = {Çakıroğlu, Ömer Faruk},
title = {Beyond the AUROC: data leakage, random splitting, and neglected calibration in critical care machine learning},
journal = {Eurasian Journal of Critical Care},
year = {2026},
volume = {8},
number = {2},
pages = {113--114},
doi = {10.55994/ejcc.1993408}
}
EUVIMED