Fast SVM-based epileptic seizure prediction employing data prefetching

Abstract

To achieve high prediction accuracy for epileptic seizure prediction, a support vector machine (SVM) has been adopted due to its robust classification performance. However, in order to use an SVM for real-time applications such as seizure prediction, the slow classification speed of an SVM should be addressed. For this purpose, data prefetching that enhances the classification speed of an SVM by mitigating the gap between the processor and the main memory is employed.