Implementation of a machine learning application in preoperative risk assessment for hip repair surgery

BMC Anesthesiology

Table 5 Predictive performance of machine learning algorithms on prolonged length-of-stay*

Algorithm	Accuracy	Sensitivity	Specificity	AUROC (95% CI)
Logistic Regression	0.745	0.742	0.745	0.831 (0.791–0.871)
Random Forest	0.778	0.783	0.778	0.854 (0.818–0.890)
SVM	0.651	0.650	0.651	0.730 (0.679–0.780)
KNN	0.643	0.625	0.644	0.681 (0.627–0.736)
lightGBM	0.773	0.767	0.774	0.853 (0.815–0.892)
MLP	0.727	0.741	0.726	0.824 (0.791–0.871)
XGBoost	0.747	0.750	0.747	0.837 (0.797–0.876)

^* Prolonged length-of-stay: hospital stay longer than that of 90 percentiles in the validated cohort. Prolonged hospital stay: 401 patients
Abbreviations: AUROC area under receiver operating characteristic curve, SVM support vector machine, KNN K nearest neighbor, light GBM light gradient boosting machine, MLP multi-layer perception, XGBoost extreme gradient boosting, CI confidence interval

ISSN: 1471-2253