INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
SLEEP DISORDER CLASSIFIERS USING EEG SIGNAL PROCESSING
Authors Name:
DHANYA N
, Dr MEENA SRINIVASAN
Unique Id:
IJSDR1703034
Published In:
Volume 2 Issue 3, March-2017
Abstract:
Any change in sleep pattern which negatively affects the health is termed as sleep disorder. This paper aims to detect different types of sleep disorders and analyze the performance of the classifiers namely neural networks, ANFIS and SVM. Sample EEG signals are taken from CAP SLEEP DATABASE. For each signal, the information from at least 3 EEG channels (F3 or F4, C3 or C4, O1 or O2) is tabulated. EEG signals consists of different frequency bands namely, alpha (8-12 Hz) and beta (14-32 Hz) theta (4-8 Hz). A band pass filter is designed to extract alpha band from the EEG dataset and average power spectral density of this band is computed using Welch method. Statistical parameters like mean, standard deviation of sample data and obtained PSD of alpha wave are fed as input to the different classifier. A FIS model using grid partitioning method is generated. System was trained for 80 samples using hybrid algorithm for 25epochs in 2 seconds and the training error was 0.66257.20 samples were used to test the classifier and the results shows an average testing error of 0.50924 and an accuracy of about 91%. The SVM testing gave an accuracy of 92%. The disorders identified and classified in this paper are bruxism,narcolepsy,insomnia and nocturnal frontal lobe epilepsy. The data is read as ‘1’ for bruxism ‘2’ for epilepsy, ‘3’ for no sleep disorder, ‘4’ for insomnia and ‘5’ for narcolepsy.
Keywords:
Keywords—EEG,ANFIS,SVM,FIS,PSD
Cite Article:
"SLEEP DISORDER CLASSIFIERS USING EEG SIGNAL PROCESSING", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.2, Issue 3, page no.223 - 228, March-2017, Available :http://www.ijsdr.org/papers/IJSDR1703034.pdf
Downloads:
000337070
Publication Details:
Published Paper ID: IJSDR1703034
Registration ID:170105
Published In: Volume 2 Issue 3, March-2017
DOI (Digital Object Identifier):
Page No: 223 - 228
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn