Table 4.

Area under the precision-recall curve (AUPRC) with 95% confidence intervals, to compare and select the best model

Models	Primary outcome		Secondary outcome
	(highest care level within 24 hours)		(care level at the 24th hour)
	Baseline AUPRC = 0.13		Baseline AUPRC = 0.09
	Full feature	Simple feature	Full feature	Simple feature
Gradient Boosting*	0.65 (0.63–0.68)	0.52 (0.50–0.56)	0.50 (0.47–0.54)	0.41 (0.39–0.45)
Random Forest	0.59 (0.57–0.62)	0.49 (0.47–0.52)	0.46 (0.42–0.49)	0.39 (0.36–0.42)
Logistic Regression (elastic net)	0.52 (0.49–0.55)	0.46 (0.44–0.49)	0.39 (0.37–0.43)	0.36 (0.34–0.40)
Feed-Forward Neural Networks	0.56 (0.53–0.59)	0.45 (0.43–0.49)	0.42 (0.39–0.45)	0.38 (0.36–0.42)

Notes: The baseline of AUPRC is equal to the fractions of positive cases for each outcome.

*Best models with highest AUPRC.