Effects of Data Augmentation with the BNNSMOTE Algorithm in Seizure Detection Using 1D-MobileNet

Algorithm 1

BNNSMOTE algorithm.

	Input: seizure sample set and nonseizure sample set
	Output: seizure sample set after imbalance treatment
(1)	Calculate the k-nearest neighbor set of
(2)	Denoised seizure sample set
(3)	Calculate the k-nearest neighbor nonseizure set of
(4)
(5)	Calculate the k-nearest neighbor seizure set of
(6)
(7)	Calculate the number of new samples that each seizure sample should generate
(8)	fordo
(9)	fordo
(10)
(11)
(12)	end
(13)	end
(14)