Teager Energy Based Filter-Bank Cepstra in EEG Classification for Seizure Detection Using Radial Basis Function Neural Network

Table 3

Percentage average accuracy of RBFNN analysis using linear, logarithmic, and Mel scale TE-FB-CEP methods (𝑊=868, 𝑁=9, and 𝑠=2) for composite vector-2 in discriminating eight classification problems (CPs).

CP	Linear scale TE-FB-CEP	Logarithmic scale TE-FB-CEP	Mel scale TE-FB-CEP
1	98.51	99.99	98.40
2	97.05	98.90	97.08
3	97.20	98.73	97.29
4	98.62	99.28	98.72
5	98.26	98.72	98.41
6	99.22	99.30	99.20
7	97.88	98.61	98.12
8	98.83	99.12	98.76