Research article

Explainable models for early sepsis detection

Yılmaz, A.; Keller, M.; Rossi, L. · Article a2

https://doi.org/10.1234/ajm.2026.012

Abstract

We present a calibrated gradient boosting pipeline for early sepsis prediction using routinely collected EHR features. The model improves AUROC while remaining interpretable for clinical triage teams.

Keywords

sepsis · machine learning · explainability · EHR

Cite this article

APA. Yılmaz, A., Keller, M., & Rossi, L. (2026). Explainable models for early sepsis detection. Anatolian Journal of Medicine, 12(2), 12–28. https://doi.org/10.1234/ajm.2026.012

MLA. Yılmaz, Ayşe, et al. “Explainable Models for Early Sepsis Detection.” Anatolian Journal of Medicine, vol. 12, no. 2, 2026, pp. 12–28.

Chicago. Yılmaz, Ayşe, Markus Keller, and Luca Rossi. 2026. “Explainable Models for Early Sepsis Detection.” Anatolian Journal of Medicine 12 (2): 12–28.