Prediction of Epileptic Seizure using Support Vector Machine and Genetic Algorithm

Abstract

Epilepsy is a condition defined by the occurrence of epileptic seizures. An epileptic seizure is a brief episode of symptoms caused by abnormal electrical activities in the brain. A common way to treat epileptic seizure is the use of medication. When medication fails, surgery is usually the proposed but surgeries have been found to fail in numerous cases leaving victims with no option than to manage their condition. This scenario, prompt the prediction of epileptic seizures earlier before its invasion so that appropriate precautions can be observed. This research proposes machine learning algorithm; support vector machine and genetic algorithm for the prediction of epileptic seizures. Genetic algorithm was adapted for feature selection while support vector machine was used in classifying EEG signals as seizure or non-seizure signals. The developed model generated accuracy 97.73%, sensitivity 97% and specificity 97%.