Epilepsy classification through multi-label dimensionality reduction through dependence maximization and elite genetic algorithm

Abstract

One of the global neurological disorders affecting the cerebral cortex of the brain is epilepsy. It is considered as a persistent neurological disorder that exists since a long period of time. Characterized by continuous, spontaneous and recurrent seizures, they cause a great harm to the patient. The activities of the neurons can be easily recorded by Electroencephalogram (EEG). A mandatory requirement for the automated detection of epilepsy using EEG holds a high impact in the diagnosis and analysis of the disorder. Since the EEG recordings occur for a long period of time, processing it is difficult and so the dimensions of it are reduced with the help of Multi-label dimensionality reduction through dependence maximization. The dimensionally reduced values are then classified with the help of Elite Genetic Algorithm (GA). The results show that an average accuracy of about 89.68%, an average time delay of about 2.44 seconds along with an average performance index of 71.76% is obtained.