Table 3 Performance comparisons of five machine learning methods.

From: Machine learning-based risk prediction model construction of difficult weaning in ICU patients with mechanical ventilation

 

Sensitivity

Recall

Specificity

Accuracy

F1

Logistic

0.709

0.969

0.769

0.714

0.812

SVM

0.782

0.878

0.676

0.755

0.827

Random forest

0.767

0.888

0.676

0.748

0.825

XG boost

0.782

0.878

0.676

0.755

0.827

Light GBM

0.733

0.898

0.630

0.714

0.807

  1. Bold indicate best performance.
Back to article page