Machine learning-based predictive model for the development of thrombolysis resistance in patients with acute ischemic stroke

Table 2 Statistical comparison of machine learning models for the occurrence of thrombolysis resistance

Model	AUC(95%CI)	sensitivity	specificity	PPV	NPV	Accuracy	Precision	Recall	F1 value
Training group
LR	0.778(0.704–0.852)	0.724	0.745	0.636	0.814	0.737	0.636	0.724	0.677
LASSO	0.789(0.717–0.862)	0.741	0.777	0.672	0.830	0.763	0.672	0.741	0.705
SVM	0.760(0.683–0.837)	0.862	0.585	0.562	0.873	0.691	0.562	0.862	0.680
XGBoost	0.783(0.710–0.855)	0.897	0.585	0.571	0.902	0.704	0.571	0.897	0.698
RF	0.782(0.707–0.856)	0.741	0.766	0.662	0.828	0.757	0.662	0.741	0.699
Testing group
LR	0.664(0.532–0.796)	0.767	0.528	0.575	0.731	0.636	0.575	0.767	0.657
LASSO	0.765(0.649–0.881)	0.767	0.694	0.676	0.781	0.727	0.676	0.767	0.717
SVM	0.756(0.638–0.874)	0.933	0.500	0.609	0.900	0.697	0.609	0.933	0.737
XGBoost	0.721(0.599–0.844)	0.800	0.583	0.615	0.778	0.682	0.615	0.800	0.695
RF	0.693(0.566–0.819)	0.800	0.528	0.585	0.760	0.652	0.585	0.800	0.558

Abbreviations LR, logistic regression; LASSO, the least absolute shrinkage and selection operator; XGBoost, extreme gradient boosting; SVM, support vector machine; RF, random forest; AUC, area under the curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value

ISSN: 1471-2377