Skip to main content

Table 2 Performance of each model for prediction

From: Machine-learning predictions for acute kidney injuries after coronary artery bypass grafting: a real-life muticenter retrospective cohort study

  

Cutoff

AUC

Sensitivity

Specificity

Positive predictive value

Negative predictive value

Internal validation group

      
 

LightGBM

0.0715

0.8027

0.7011

0.7889

0.3739

0.9790

 

SVM

0.2000

0.7805

0.8197

0.6111

0.3957

0.9161

 

Softmax

0.1192

0.7568

0.6459

0.7667

0.2918

0.9319

 

RF

0.1754

0.7292

0.7189

0.6222

0.2995

0.9079

External validation group of QLH

      
 

LightGBM

0.0916

0.8798

0.8354

0.8000

0.3025

0.9791

 

SVM

0.1409

0.8819

0.9113

0.7037

0.4113

0.9718

 

Softmax

0.0954

0.8411

0.8414

0.6791

0.3588

0.9704

 

RF

0.1592

0.7861

0.8916

0.6000

0.4239

0.9611

External validation group of GHN

      
 

LightGBM

0.0795

0.7801

0.7759

0.6909

0.2138

0.9070

 

SVM

0.1331

0.7504

0.7874

0.6545

0.3302

0.9327

 

Softmax

0.0600

0.6941

0.5776

0.7091

0.2806

0.9394

 

RF

0.1932

0.6777

0.8649

0.4182

0.4600

0.9039

  1. Abbreviation: AUC, area under the curve; SVM, support vector machine; RF, random forest