Predicting outcomes in older ED patients with influenza in real time using a big data-driven and machine learning approach to the hospital information system

Table 3 Evaluation report using the best model with the SMOTE preprocessing algorithm on the outcomes of testing dataset of older ED patients with influenza

Outcome	Number	Negative outcome	Positive outcome	Accuracy	Sensitivity	Specificity	PPV	NPV	AUC
Hospitalization (random forest)	5508	2901	2607	0.769	0.744	0.791	0.762	0.775	0.840
Pneumonia (random forest)	5508	3431	2077	0.679	0.681	0.679	0.562	0.778	0.765
Sepsis or septic shock (random forest)	5508	5201	307	0.795	0.750	0.798	0.179	0.982	0.857
ICU admission (XGBoost)	5508	5449	59	0.912	0.722	0.914	0.085	0.997	0.902
In-hospital mortality (logistic regression)	5508	5387	121	0.816	0.806	0.816	0.089	0.995	0.889

SMOTE Synthetic minority oversampling technique; ED Emergency department; PPV Positive predictive value; NPV Negative predictive value; AUC Area under the curve; ICU Intensive care unit; XGBoost Extreme Gradient Boosting

ISSN: 1471-2318