Comparison of artificial neural network and support vector machine classification methods in diagnosis of migraine by using EEG

%23 of human has a migraine disease which is a painful and throbbing brain disorders. Cause of migraine isn´t known and accepted automatic diagnose method of migraine by biomedical equipment isn´t available yet. But researches are continuing for diagnose of migraine by assistance of triggering factors like flash stimulation. To obtain information about migraine generally change of EEG signals under flash stimulation is used as a method. In this study aim is performance analysis of classification methods of EEG signals obtained from stimulated migraine patient by flash light for automatic migraine diagnose. Firstly EEG signals obtained from both migraine patients and healthy subjects are transformed to frequency domain by using (AR) Burg method. And these frequency spectrums are classified by using artificial neural network and support vector machine classification algorithm. According to these classification results which classification algorithm has a better performance for migraine diagnose is determined.