Utilizing imbalanced electronic health records to predict acute kidney injury by ensemble learning and time series model

Table 2 Performance of imbalanced learning techniques on ICUC

	AUC		Sensitivity		F1-socre		AP
	24h	48h	24h	48h	24h	48h	24h	48h
Random Undersample	0.81	0.78	0.75	0.68	0.58	0.44	0.59	0.41
Random Oversample	0.78	0.69	0.64	0.43	0.62	0.44	0.66	0.46
Cost-sensitive XGBoost	0.78	0.70	0.64	0.45	0.61	0.46	0.67	0.47

ISSN: 1472-6947