Binary and multiclass classifiers based on multitaper spectral features for epilepsy detection

Epilepsy is one of the most common neurological disorders that can be diagnosed by means electroencephalogram (EEG) analysis, in which following epileptic events observed: pre-ictal, ictal, post-ictal, and interictal. In this paper, we present a novel method for epilepsy detection employing binary multiclass classifiers. For feature extraction, total 105 measurements were extracted from power spectrum, spectrogram, bispectrogram. classifier building, widely known machine learning algorithms used. Our was applied publicly available EEG database. As result, BP-MLP (backpropagation based on multilayer perceptron) SMO_Pol (sequential minimal optimization supported polynomial kernel) reached highest accuracy (100%) (98%) classification problems. Subsequently, statistical tests did not find better performance model. evaluation confusion matrices, it also impossible to identify stands out concerning other models classification. comparison related words, our predictive competitive results.