Research Article

On the Use of Wavelet Domain and Machine Learning for the Analysis of Epileptic Seizure Detection from EEG Signals

Table 6

Comparison with the existing state of the art.

ReferencesYearMethodsCasesCA (%)

[30]2006DWT adaptive neurofuzzy networkAD-E85.9
[31]2009DWT + ApEn and surrogate data analysisACD-E96.65
[18]2010Line length feature and ANNA-E99.6
ACD-E97.75
ABCD-E97.5
[20]2011Statistical features from DWT + kNN classifierA-E100.0
AB-CD-E95.6
[21]2012Permutation entropy + SVMA-E100.0
B-E82.88
C-E88.0
D-E79.94
[32]2012Statistical features from DWT + PCA + ANN classifierA-E100.0
[17]2012ApEp + SampEp + phase entropy 1 + phase entropy 2-Fuzzy Sugeno classifierAB-CD-E98.1
[29]2013DWT + permutation and sample entropy + Hurst exponent + genetic algorithm + extreme learning machine (ELM)A-E94.8
[34]2013DWT + Hurst and Lyapunov exponentB-E96.5
[24]2014Dual-tree complex wavelet transform + kNN transformA-E100.0
CD-E100.0
ABCD-E100.0
[25]2015Empirical mode decomposition-based temporal spectral features + SVMA-E100.0
[26]2015DTCWT + complex-valued neural networkAB-CD-E98.28
[33]2016DTCWT + general regression neural networkA-E100.0
B-E98.9
C-E98.7
D-E93.3
AB-E99.2
[28]2016Key-point based local binary pattern of EEG signalsCD-E99.45
AB-CD-E98.8
[35]2017Tunable-Q wavelet transform + kNN entropy + SVMAB-CD-E98.6
[4]2016DWT – MVP + SD + AVP – NB/kNN classifierA-E100.0
B-E99.25
C-E99.5
D-E95.62
AB-E99.16
AC-E99.5
AD-E97.08
BC-E98.25
BD-E96.5
CD-E98.75
ABC-E98.68
ACD-E97.31
BCD-E97.72
ABCD-E97.1
[36]2018CNNAB-CD-E88.67
[37]2018CNNA-E100
B-E99.8
C-E99.1
D-E99.4
AB-E99.8
AC-E99.7
BC-E99.5
BD-E99.6
CD-E99.7
ABC-E99.97
ACD-E99.8
BCD-E99.3
ABCD-E99.7
AB-CD99.9
AB-CD-E99.1
AB-CDE99.7
This work (University of Bonn dataset)2020DWT + SVM/KNN/NB/DTA-E100.0
B-E100.0
C-E100.0
D-E97.5
AB-E100.0
AC-E98.67
AD-E98.0
BC-E98.67
BD-E96.33
CD-E98.0
ABC-E99.0
ABD-E97.0
ACD-E98.25
BCD-E97.0
ABCD-E100.0
AB-CD82.5
AB-CD-E95.0
This work (real-time clinical dataset)2020DWT + SVM/KNN/NB/DTHealthy- epileptic patient91.67