Seizure Type Detection in Epileptic EEG Signal using Empirical Mode Decomposition and Support Vector Machine

Epilepsy is a serious neurological disorder that needs more attention by society. The International League Against Epilepsy (ILAE) mentioned that the term epilepsy referred to the number of seizure occurred in patients. Electroencephalogram signal is a common epilepsy diagnostic tools used by the neurologist. Research about the detection and classification of the epileptic signal from the EEG signal has massively conducted. In this research, we detect and classify four types of seizures which are focal non-specific seizure (FNSZ), generalized non-specific seizure (GNSZ), simple partial seizure (SPSZ), and tonic-clonic seizure (TNSZ). The EEG signal used was taken from Temple University Hospital EEG Seizure Corpus (TUSZ) version 1.2.0. The EEG signal decomposed with empirical mode decomposition (EMD) to extract five levels of intrinsic mode functions (IMFs). Feature extraction is done by calculating the mean, variance, skewness, kurtosis, standard deviation, and interquartile range. Support Vector Machine (SVM) used for classification with five-fold cross-validation. The best accuracy obtained is 95% by using quadratic SVM kernel.