Real-time AI prediction for major adverse cardiac events in emergency department patients with chest pain

Table 3 Comparisons of predictive accuracies among random forest, logistic regression, SVC, and KNN models for adverse outcomes of ED patients with chest pain

Outcomes and predictive models	Accuracy	Precision	Sensitivity	Specificity	F1	AUC
AMI < 1 month
Random forest	0.915	0.916	0.915	0.882	0.915	0.915
Logistic regression	0.868	0.885	0.868	0.766	0.867	0.868
SVC	0.631	0.635	0.631	0.538	0.627	0.631
KNN	0.865	0.880	0.865	0.766	0.864	0.865
All-cause mortality < 1 month
Random forest	0.999	0.999	0.999	1.000	0.999	0.999
Logistic regression	0.716	0.717	0.716	0.690	0.716	0.716
SVC	0.656	0.660	0.656	0.584	0.654	0.656
KNN	0.969	0.971	0.969	0.940	0.969	0.969

SVC support-vector clustering; KNN K-nearest neighbors; ED emergency department; F1 2 x (precision x recall/precision + recall); AUC area under the curve; AMI acute myocardial infarction

ISSN: 1757-7241