BMC Medical Informatics and Decision Making

Table 6 Performance of derived ETSM on ICUC in the experiment of predicting AKI 48 hours ahead

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

Model	AUC	Sensitivity	F1-score	AP
	(95% CI)	(95% CI)	(95% CI) 7 (95% CI)
ETSM	0.776 ±0.002	0.683 ±0.004	0.437 ±0.003	0.406 ±0.004
ETSM-ex	0.775 ±0.003	0.684 ±0.006	0.434 ±0.003	0.396 ±0.005*
ETSM-bool	0.786 ±0.003	0.702 ±0.005	0.453 ±0.003	0.434 ±0.006
ETSM-times	0.806 ±0.002	0.739 ±0.006	0.476 ±0.003	0.476 ±0.005

Note: CI = confident interval
^*indicates ETSM significantly outperforms the baseline with p <0.01 using Student t-test

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com