Analysis of Electroencephalogram Signals in Epileptic Seizure Recognization

Abstract

SVM approach in machine learning is utilised in the exploration of Electroencephalography signals aimed at seizure disorder recognition. The nonlinear subtleties in the unique Electroencephalography are computed in the form of the Hurst (H) and classified using SVM. The method of Electroencephalography examination contains of two phases, explicitly the pre-processing and Feature extraction analysis. The synthetic thus created decision making system is familiar with classification of the signals which are available in brain, as whether it is a primary seizure or secondary seizure. EEG may be a brain signal processing technique that allows understanding the complex inner mechanisms of the brain and abnormal brain waves. The analysis of brain waves plays an important role within the diagnosis of varied brain disorders. The classification capability of the Hurst measures is verified by Support vector machine classification. By employing a target EEG signal, and pre-processing and also by performing Feature extraction desirable outcomes is obtained.