Skip to main content

Table 5 Mean testing accuracy and the AUC of ensemble learning methods after 50 replicates with standard deviation in the brackets

From: Ensemble machine learning approach for screening of coronary heart disease based on echocardiography and risk factors

Model

Accuracy

AUC

Two-step stacking (14 models)

\(\mathbf {87.7}\% \mathbf {(0.023)}\)

0.904 (0.026)

Two-step stacking (3 models)

\(79.4\% (0.028)\)

0.822 (0.030)

Traditional stacking (14 models)

\(81.8\% (0.033)\)

0.854 (0.034)

Traditional stacking (3 models)

\(76.7\% (0.038)\)

0.798 (0.037)

Weighted voting (14 models)

\(73.3\% (0.033)\)

0.751 (0.040)

Weighted voting (3 models)

\(71.7\% (0.035)\)

0.728 (0.037)

Two-step stacking with GLPS only

\(63.3\% (0.034)\)

0.674 (0.047)

  1. Best results are bolded