Stacking ensemble of RF, LightGBM, and DNN achieves NPV greater than 99.9% on internal test sets for meningitis detection in ICU despite class imbalance, with robust sensitivity on external eICU data.
Risk Features We analyzed feature importance to improve data quality and increase diagnostic difficulty, emulating the scenes in the ER
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Ensemble Deep Learning Models for Early Detection of Meningitis in ICU: Multi-center Study
Stacking ensemble of RF, LightGBM, and DNN achieves NPV greater than 99.9% on internal test sets for meningitis detection in ICU despite class imbalance, with robust sensitivity on external eICU data.