Table 3 Comparison of the performance of all the classifiers implemented.

From: Robust diabetic prediction using ensemble machine learning models with synthetic minority over-sampling technique

Classifier

AUC (Avg ± Std)

Accuracy (Avg ± Std)

Sensitivity

Precision

Specificity

False Omission Rate

Diagnostic Odds Ratio

K-Nearest Neighbor

0.898 ± 0.035

0.881 ± 0.025

0.898

0.859

0.866

0.097

56.850

Decision Tree

0.938 ± 0.007

0.867 ± 0.011

0.868

0.855

0.866

0.122

42.512

Naïve Bayes

0.867 ± 0.023

0.803 ± 0.022

0.796

0.790

0.810

0.184

16.634

Random Forest

0.963 ± 0.009

0.875 ± 0.018

0.883

0.883

0.893

0.107

63.140

AdaBoost

0.950 ± 0.007

0.865 ± 0.013

0.855

0.860

0.873

0.131

40.657

XGBoost

0.962 ± 0.012

0.901 ± 0.016

0.895

0.895

0.906

0.094

82.686

AdaBoost + XGBoost

0.968 ± 0.015

0.904 ± 0.023

0.897

0.902

0.911

0.093

89.108

Back to article page