Investigation of Wavelet Transform Performance in Classification of Vigilance States

In this study, we worked on a method of analysis of EEG signals classification using artificial neural networks after evaluating differences in alert, drowsy and sleeping conditions (observed in both time domain and in time-scale domain we got though wavelet transform). As it was understood from the discoveries´ graphics which were got, it had been supplied for neural network in using classification problem successfully. Neural networks performance has been changing according to the changing of neural networks learning coefficient, activation function values, hidden layer number and hidden layer neuron number. These values had been made optimal according to experimental results