Epileptic Seizure Detection with Hybrid Time-Frequency EEG Input: A Deep Learning Approach

Table 3

Performance of normal vs. seizure case (A vs. E) with 20-fold cross-validation.

Methods	K1	K2	K3	K4	K5	K6	K7	K8	K9	K10	K11	K12
EEG	0.9556	0.9667	0.9778	0.9278	0.9611	0.9167	0.9667	0.9556	0.9444	0.9611	0.9536	0.9737
DWT	0.9278	0.9389	0.9833	0.9611	0.9889	0.9833	0.9667	0.9833	0.9444	0.9667	0.9579	0.9579
FFT	0.9222	0.9889	0.9889	0.9722	0.9667	1.0	0.9833	0.8944	0.8556	0.9944	0.9895	0.9895
STFT	0.9222	0.9778	0.95	0.9222	1.0	0.9889	0.9389	1.0	0.9778	0.9667	0.9158	0.9789
Hybrid	0.9667	0.9667	0.9944	0.9722	0.9778	0.9611	0.9833	0.9944	0.9944	0.9833	0.9579	0.9579
Methods	K13	K14	K15	K16	K17	K18	K19	K20	Mean	Variance
EEG	0.9737	0.9684	0.9474	0.9947	0.8842	0.9684	0.9474	0.9895	0.9498	0.0297
DWT	0.9947	0.9368	0.9895	0.9579	0.9474	0.9947	0.9579	0.9895	0.9595	0.0292
FFT	0.9895	0.9947	0.9947	0.9789	0.9947	1.0	0.9947	0.9842	0.9371	0.12
STFT	0.8368	0.7895	0.9632	1.0	0.9053	0.8947	0.9947	0.9684	0.9431	0.0564
Hybrid	0.9895	0.9789	0.9526	0.9947	0.9421	0.9736	0.9684	0.9789	0.9639	0.0267