Predicting venous thromboembolism in hospitalized trauma patients: a combination of the Caprini score and data-driven machine learning model

Table 2 Prediction Metrics of Different Models based on EHR Data (Compared with the Caprini RAM)

Method	No. of Selected features	Accuracy	Precision	TPR_c (Recall)	FPR_c	AUC
Lasso + RF	26	0.636	0.332	0.692	0.320	0.759
Ridge + RF	52	0.620	0.324	0.707	0.340	0.760
ElasticNet + RF	29	0.635	0.331	0.685	0.310	0.763
LR + RF	14	0.619	0.329	0.718	0.330	0.769
MIE + RF	42	0.642	0.334	0.684	0.310	0.759
Caprini RAM	–	0.750	0.444	0.667	0.227	0.773

ISSN: 1471-227X