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

Algorithm 1

Algorithm of modified-SMOTE.

	Input:
	(1) Data of minority class , where
	(2) size of minority instances (T)
	(3) Augmentation in percentage (N)
	(4) size of nearest neighbors (k)
	Output: Augmented Data
	for eachdo
	(1) Find out the k-nearest neighbour from minority dataset of
	(2)
	end
	foreachdo
	(1) Choose the one of the k-nearest neighbour said
	(2) Compute difference:
	(3) Compute gap:
	(4) Synthetic:
	(5) Append to
	(6) decrements of
	end
	Note: The derivation of the gap has been proved in Appendix A