Sakarya University Journal of Computer and Information Sciences
Yazarlar: Mustafa TOSUN, Mustafa ERGİNLİ, Ömer KASIM, Burak UĞRAŞ, Şems TANRIVERDİ, Tayfun KAVAK
Konular:Bilgisayar Bilimleri, Bilgi Sistemleri
DOI:10.35377/saucis.01.02.443999
Anahtar Kelimeler:EEG,Backpropagation neural network,Welch method,Signal processing
Özet: In recent years, as a result of the technological development, there has been a significance improvement on the computer interface. Electroencephalogram (EEG) signals are widely used in Brain Computer Interface (BCI) methods. By using EEG data, the imagination of movement with physical motion can be classified. In this study, EEG data of a 21-years-old man who used his right hand and who didn’t show any disease symptom was used. Part of this EEG data demonstrates the recordings of forward and backward movement of the left and right hand. The other data indicates the records of imagination of motion without any physical movement. Using the Welch method, the power densities of the frequencies of 1-48 Hz of the EEG data were calculated. The obtained data sets were applied to the designed Back Propagation Neural Network (BPNN). At the end of the network training, the Mean Squared Error (MSE) value of 4.6731x10-23 has been reached. When the test data set, which consists of imaginary and motion data is applied to the trained network, imagination and motion data are classified with accuracy of 99.9975%.
Dergi editörleri editör girişini kullanarak sisteme giriş yapabilirler. Editör girişi için tıklayınız.