Skip to main content

Table 3 Prediction performance using percentage split

From: Predicting diabetic retinopathy and identifying interpretable biomedical features using machine learning algorithms

Model

Training

Test

AUC

Acc.

Sens.

Spec.

AUC

Acc.

Sens.

Spec.

SVM

0.783

0.708

0.787

0.664

0.839

0.795

0.933

0.724

LR

0.749

0.679

0.703

0.660

0.802

0.727

0.813

0.679

ANN

0.875

0.762

0.756

0.768

0.777

0.682

0.682

0.682

DT

0.719

0.685

0.660

0.718

0.768

0.727

0.708

0.750

  1. AUC, accuracy, sensitivity, and specificity of different machine learning algorithms using training (i.e., 80%) and test (i.e., 20%) data sets are evaluated
  2. aBest evaluation measures in test set are underlined