Supervised Machine Learning Based Noninvasive Prediction of Atrial Flutter Mechanism from P-to-P Interval Variability under Imbalanced Dataset Conditions

Table 10

Performance evaluation by linear classifiers at 400% data augmentation with five-fold cross-validation.

Technique	Performance	LDA	LOG	SVM
Modified-SMOTE	Accuracy	77.81	76.88	77.45
	Specificity	41.35	49.50	36.25
	Sensitivity	95.60	90.24	97.56