BMC Medical Informatics and Decision Making

Table 4 Comparison of predictive validity of the optimal ML models, proposed model and Padua

From: A new risk assessment model of venous thromboembolism by considering fuzzy population

Model Name	Training set			Test set
Model Name	Sensitivity	Specificity	Youden	Sensitivity	Specificity
SVM	0.8042	0.7511	0.5554	0.8292	0.7104
RF	0.8307	0.7975	0.6282	0.8780	0.7643
GBDT	0.8148	0.8223	0.6372	0.8780	0.7787
LR	0.7725	0.7864	0.5589	0.8537	0.7708
XGBoost	0.7883	0.8302	0.6185	0.8537	0.7919
Padua	0.8466	0.6127	0.4593	0.9024	0.6330
Proposed method	0.8995	0.6786	0.5781	0.9024	0.6481

The optimal ML models and proposed model were selected by maximizing the value of Youden index on training data. Note that metrics of predictive validity on training process were computed using all patients from training data

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com