Multiscale Hjorth Descriptor on Epileptic EEG Classification

Table 1

Comparison with other studies.

Author	Method	Number of features	Scenarios	Accuracy (%)
Orhan et al. [34]	DWT, K-mean clustering	56	ZO-NF-S	95.6
		56	ZONF-S	99.6
		18	Z-F-S	96.7
Bhattacharyya et al. [5]	Tunable Q-factor wavelet transform, multiscale entropy	16	ZONF-S	99
		16	ZO-NF-S	98.6
Dhar et al. [33]	Cross wavelet transform, PNN	42	ZO-NF-S	98.2
	Cross wavelet transform, SVM	42	ZO-NF-S	94
	Cross wavelet transform, LVQ	42	ZO-NF-S	97.5
Nicolaou et al. [6]	Permutation entropy, SVM	1	ZO-NF-S	98.6
		1	ZONF-S	86.1
Li et al. [4]	Wavelet-based nonlinear analysis, SVM	24	ZO-NF-S	99.3
		24	Z-F-S	99.6
		24	ZONF-S	99.4
Wijayanto et al. [35]	Higuchi, Katz FD, SVM	2	ZO-NF-S	96.6
		2	ZONF-S	98.4
		2	Z-N-S	98.7
		2	Z-F-S	97.7
		2	O-N-S	98.7
		2	O-F-S	97.7
Yazid et al. [36]	Local binary pattern transition histogram (LBPTH) and local binary pattern mean absolute deviation (LBPMAD), SVM, K-NN	18	ZONF-S	99.74
Proposed method	Multiscale hjorth descriptor, SVM	15	ZO-NF-S	97.6
		15	ZONF-S	99
		15	Z-N-S	97.3
		15	Z-F-S	97.7
		15	O-N-S	96.7
		15	O-F-S	99.3