The Study Of Imbalance Data Processing And EEG Feature Extraction

Epilepsy is one of the most common neurological disorders,which is characterized by recurrent and transient epileptic seizures.It may increase the risk of sustaining physical injuries and even result in death.The accurate seizure detection has been considered to be the most important step for the diagnosis and treatment follow-up in epilepsy patients.However,the traditional visual examination of long-term and imbalanced EEGs by a trained neurologist is a time-consuming and subjective process.Therefore,there has been an increasing interest in the study of the automated seizure detection using EEGs in recent years.In order to realize it successfully,how to design an appropriate data balance method to preprocess the long-term EEG,as well as how to design an appropriate feature extraction method,are then two important topics in the study.This paper proposes a novel data balance method K-M-S,as well as a new kernel-radius-based feature extraction method.The contents of the paper are presented as follows:Chapter 1 systematically introduces the background and development of automatic seizure detection using EEGs.Chapter 2 briefly presents some preliminaries about epilepsy,EEG,EEG databases as well as classifiers in the study of automatic seizure detection.Chapter 3 proposes a novel data balance method K-S-M,which is based on the k-means clustering method,Silhouette index and M-neighboring under-sampling.The experimental verification is then shown.Chapter 4 designs a new kernel-radius-based feature extraction method from the perspective of nonlinear dynamical analysis.The experimental verification is then shown.