Dalgacik Donfiifim Performansinin Incelenmesi Investigation ofWavelet Transform Performance in Classifi'cation ofVigilance States
Hatice Batar, A. Hamit Suba · 2006
sinyali zamanalanindan frekans alanina donu,sturup frekans Inthisstudy, weworkedonamethodofanalysis ofEEG bile,enleri onemli oldugu zamansik,a ba,vurulan bir signals classification usingartificial neural networks after tekniktir. Fourier analizinde, frekans alanina ge,ildiginde evaluating differences inalert, drowsy andsleeping conditionszamanalani kaybolur. Duragan sinyallerde fourier analizi iyi (observed inbothtimedomain andintime-scale domain we biryontem olmaktadir, fakat EEGi,aretleri duragan olmayip gotthough wavelet transform). Asitwasunderstood fromthe sulrekli degi,ir. Bui,aretlerde aradasirda kisasulreli ortaya discoveries' graphics whichweregot, ithadbeensupplied for ,ikan dikdarbeler vekompleks dalgalar te,his a,isindan neural network inusing classification problem successfully. onemli bilgiler ta*ir. Budurumda buozel spektral bile,enlerin Neural networks performance hasbeenchanging according to hangizamanaraliginda meydanageldigi onemli olabilir ve thechanging of neuralnetworks learning coefficient, fourier analizi yetersiz kalir [2]. activation function values, hidden layer numberandhidden