Optimization of Sleep Stage Classification using Single-Channel EEG Signals

Classification of various stages of sleep is mandatory for the diagnosis and treatment of sleep disorders. Manual scoring is a time-consuming and tedious task as well as it requires sleep specialists. Therefore, automatic sleep stage classification is necessary. In this paper, we have utilized state-of-the-art signal processing and machine learning techniques for sleep stage classification using single-channel EEG signal. Three cases of sleep classification have been done using support vector classifier (SVC), Decision tree (DT), Random forest (RF) and XGBoost (XGB). The features extracted from pre-procesed EEG have been applied to Spectral Regression dimensionality reduction technique to reduce the model complexity. The Bayesian Optimization (BO) technique is applied to optimize the hyper-parameters of the classifiers. Our proposed classification techniques provide the minimum error of 25.52%, 14.03%, and 4.93% for case I, case II and case III, respectively.