An Epilepsy Detection Method Using Multiview Clustering Algorithm and Deep Features

Table 1

Common EEG data feature.

Name	Details
Time domain feature	The EEG signal is a time series signal that changes with time. The discrete points in the signal represent the energy intensity at a certain moment or the voltage value measured at that moment. In this study, the original EEG signal is directly used as the feature data from the time domain perspective.
Spectrum feature	The EEG signal is mainly divided into 6 frequency intervals, namely, (0-2 Hz), (2-4 Hz), (4-8 Hz), (8-15 Hz), (15-30 Hz), and (30-60 Hz). The spectral features of seizures are mainly distributed between 4 Hz and 30 Hz. For the initial EEG signal, the feature data under the spectral view were obtained by Fourier transform.
Time-frequency feature	Wavelet packet decomposition (WPD) is performed on the time-domain signal to obtain feature data from a time-frequency view. The sampling interval of the large frequency domain is set at 2 Hz, and the decomposition level of the wavelet transform is 6 layers. The time-frequency feature selection interval is 4 Hz-30 Hz.