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

Algorithm 2

Feature scoring algorithm.

	Data: Set of features
	Set all score counters to 0 and of augmentation techniques (N)
	foreachdo
	feature scoring wrapper approach
	foreachdo
	Set
	Perform a wrapper evaluation of all possible
	feature combination of length l and
	calculate the performance metric
	Determine the features participating in the best
	combinations of length l and store them
	foreachdo
	Increment
	end
	end
	foreachdo
	end
	feature scoring for filter method
	Set
	foreachdo
	Determine the feature which lies within the and store them in
	end
	end