Diagnosis and Classification of Epileptic Seizure a Neurological Disorder Using Electroencephalography

摘要

An epileptic seizure is a neurological disorder which is result of sudden excessive electrical discharge from neurons which may cause loss of consciousness. The brain signals can be measured by using Electroencephalography (EEG). In this paper we analyze the EEG signal in time frequency domain and classify the signal as seizure and non-seizure. The available standard online database is used which is acquired by International standard 10–20 EEG placement system. The signal is then preprocessed to remove power noise and eye blink artifact. The features such as mean, standard deviation, variance, skewness and kurtosis are found, which are classified by classifier such as Support Vector Machine, K-Nearest Neighbor algorithm and Probabilistic Neural Network. The performances of above classifier are evaluated on the bases of sensitivity, specificity and accuracy.