Skip to main content
. 2022 Jun 17;9:854287. doi: 10.3389/fcvm.2022.854287

Table 3.

Comparison of clinical effectiveness of models.

Model Pt (%) Net income Model
net
income 		Advantages
of the
model#
Treat all Prediction
model
FLR-L1-LR 5 0.051 0.066 0.015 29
10 −0.002 0.049 0.051 46
11a −0.013 0.048 0.061 49
FLR-SVM 5 0.051 0.065 0.014 27
10 −0.002 0.048 0.050 45
11a −0.013 0.045 0.058 47
Lasso-
AdaBoost		 5 0.051 0.063 0.012 23
10 −0.002 0.045 0.047 43
11a −0.013 0.043 0.056 46
FLR-RF 5 0.051 0.064 0.013 25
10 −0.002 0.046 0.048 43
8a 0.02 0.053 0.033 38
#

The value was calculated as: (net benefit of the model– net benefit of treat all)/[pt/(1 – pt)] × 100.

a

Select the optimal threshold probability of each model according to AUC.

Pt, Threshold probability; Lasso-AdaBoost, AdaBoost with Lasso regression; FLR-L1-LR, L1 regularized Logistic regression with forward Partial Likelihood Estimation; FLR-RF, random forest with forward Partial Likelihood Estimation; FLR-SVM, support vector machine with forward Partial Likelihood Estimation.