Detection of epileptic seizure from EEG signals by using recurrence quantification analysis

Author

Kutlu, F. ; Kose, C.

fYear

2014

fDate

23-25 April 2014

Firstpage

1387

Lastpage

1390

Abstract

The pre-diagnosis of diseases with computerized systems is widely used in recent years for reducing diagnosis time and ratio of misdiagnosis. In this study, a pre-diagnosis system has been proposed which separates of healthy and epileptic seizures periods. For the experiments, EEG signals acquired from healthy and epileptic individuals were used. In feature extraction stage, recurrence quantification analysis (RQA); in classification stage, support vector machines (SVM), multilayer perceptron neural networks (MLPNN) and Naive Bayes classifiers have been utilized. Accordingly, in case of using MLPNN, 96.67% classification performance was obtained.

Keywords

Bayes methods; electroencephalography; feature extraction; medical diagnostic computing; medical signal processing; patient diagnosis; perceptrons; recurrent neural nets; signal classification; support vector machines; EEG signals; MLPNN; Naive-Bayes classifiers; RQA; SVM; classification stage; computerized systems; diagnosis time; disease prediagnosis; epileptic seizures periods; feature extraction stage; misdiagnosis; multilayer perceptron neural networks; prediagnosis system; recurrence quantification analysis; support vector machines; Artificial neural networks; Brain; Conferences; Educational institutions; Electroencephalography; Epilepsy; Signal processing; classifiers; epilepsy; epileptic seizure; recurrence quantification analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing and Communications Applications Conference (SIU), 2014 22nd

Conference_Location

Trabzon

Type

conf

DOI

10.1109/SIU.2014.6830497

Filename

6830497