Classification of sleep disorders based on EEG signals by using feature extraction techniques with KNN classifier

Electroencephalogram (EEG) is a useful biomedical signal in detection of sleep disorders. Paper discusses the human brain signal activity linked with specific sleep disorders. Objective of the work is detection and classification of sleep disorders using time-frequency analysis of EEG signal. In this paper, seven different sleep disorders along with one healthy subject have been studied. Large amount of EEG records have been taken from PhysioNet database and analyzed with the help of discrete wavelet transform (DWT). The various statistical measures like maximum value, minimum value, mean value and standard deviation value of DWT sub-bands are used as extracted features for detection of different sleep disorders. For reducing the dimension of extracted features different feature reduction techniques like principal component analysis (PCA), independent component analysis (ICA) and linear discriminant analysis (LDA) are used. The results of these feature reduction techniques are used to classify different sleep disorders using k-nearest neighbor (KNN) classifier. The performance of these classification processes is evaluated by using their accuracy to predict the sleep disorders.