BMC Medical Informatics and Decision Making

Table 5 Model performance for predicting the risk of severely injured (NISS > 15) in the hold-out analysis

From: On Scene Injury Severity Prediction (OSISP) model for trauma developed using the Swedish Trauma Registry

Metric	Case	LR	RF	XGBoost	SVM	ANN
Accuracy	1	83	83	84	84	83
Accuracy	2	91	89	91	91	89
F1-Score	1	41	43	42	34	48
F1-Score	2	44	41	46	39	43
AUC	1	0.83	0.80	0.84	0.81	0.83
AUC	2	0.89	0.85	0.89	0.85	0.87
AUCPR	1	0.54	0.47	0.54	0.54	0.53
AUCPR	2	0.54	0.43	0.53	0.53	0.46

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com