Table 3 Performance metrics of predictive models: responders vs. Nonresponders.

From: Machine learning predicts spinal cord stimulation surgery outcomes and reveals novel neural markers for chronic pain

Model

EEG features (n)

Accuracy (%)

F1 Score

AUC-ROC

Decision Tree

12

88.2

0.857

0.879

Random Forest

15

35.3

0.154

0.700

XGBoost

5

17.6

0

0.857

Logistic Regression

19

35.3

0

0.843

SVM

5

58.8

0

0.643

  1. AUC-ROC area under the curve of receiver operating characteristic, SVM support vector machine, n optimal number of EEG features.
Back to article page