Detection of Epileptic Seizures using Machine Learning

Abstract

Electroencephalogram (EEG) contains vital physiological information that provides important information about the human brain activity, which makes it of primary importance for the diagnosis and detection of epileptic seizures. According to experts, before a seizure, there is some abnormal activity in the brain called the preictal state and the challenging part is to distinguish preictal and interictal state of the brain. For such challenges, there is a need of automated models for detecting massive raw data and accurately classifying the data with low false positives. These models will help the patients as well as assist the medical team for accurate and time efficient detection. The right combination of data preprocessing methodology, feature extraction and classification will yield a higher accuracy, sensitivity and specificity resulting in accurate detection of epileptic seizures. In this research, the aim is to review different AI approaches and techniques that were used in previous research, for the detection of epilepticseizures. After review and analysis, the study aims at performing a comparative analysis on the machine learning algorithms and the bestperforming algorithms will be filtered out using Principal Component Analysis (PCA) method. Thefiltered algorithms will then finally be enhanced foraccurate detection of epileptic seizures.