Data subdivision approach enhances machine learning-based mortality prediction in pediatric ICU patients

Wenqian Chen, Benjamin Lee, Zexi Zang, Junfeng Li, Lingna Huang, Hang Xing, Yanli Ren · 2026

Read Data subdivision approach enhances machine learning-based mortality prediction in pediatric ICU patients by Wenqian Chen, Benjamin Lee, Zexi Zang, Junfeng Li, Lingna Huang, Hang Xing, Yanli Ren (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0349772.txt.

Description

Predicting which children in intensive care will die helps teams act early, but most patients survive, so the rare deaths are hard for algorithms to learn. Wenqian Chen and colleagues tackled that imbalance with blood test data from more than eight thousand young patients, in this 2026 PLOS One study.

Twenty-two first-day blood measures were used, with missing values filled by repeated estimation. Splitting the data into more balanced subgroups, and a stacked combination of models, both lifted accuracy above single models. Artificially boosting the rare cases during training helped every method. The authors say subgroup training is a practical trick for skewed medical data.

Catalog record

CREATORWenqian Chen, Benjamin Lee, Zexi Zang, Junfeng Li, Lingna Huang, Hang Xing, Yanli Ren
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37101
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0349772

Browse the full library · About the Tower of Babel