Table 1 Performance of the models in the test set.

From: Prediction of hearing recovery in unilateral sudden sensorineural hearing loss using artificial intelligence

Model

AUROC [95% CI]

AUPRC [95% CI]

Brier score

BACC

F1 score

P value†

Baseline model

LogReg

0.813 [0.737–0.889]

0.786 [0.683–0.881]

0.176

0.692

0.667

 

Machine learning models

XGBoost

0.876 [0.812–0.939]

0.830 [0.707–0.933]

0.143

0.781

0.762

0.0242

SVM

0.892 [0.835–0.949]

0.862 [0.774–0.946]

0.135

0.768

0.740

0.0705

LightGBM

0.915 [0.864–0.967]

0.890 [0.814–0.962]

0.117

0.840

0.824

0.0015*

MLP

0.911 [0.859–0.964]

0.897 [0.818–0.958]

0.120

0.823

0.804

0.0010*

  1. AUPRC area under the precision-recall curve, AUROC area under the receiver operating characteristic curve, BACC balanced accuracy, LogReg logistic regression, MLP multilayer perceptron, SVM support vector machine.
  2. †AUROC comparison to logistic regression.
  3. *P < 0.005.
Back to article page