Sleep stage based sleep disorder detection using single-channel electroencephalogram

The use of artificial intelligence in healthcare is the next generation of healthcare which can be the bridge between physician and patient. Sleep disorders hamper one's performance and are considered as a serious problem to overcome. There are several sleep disorders like excessive sleep (hypersomnia), inability to sleep (insomnia), snoring, apnea, etc. are detected with the analysis of polysomnogram (PSG) signals. Considering many signals for PSG increases the system's memory and computation requirements. Hence, in this work a machine learning model is proposed, considering single-channel electroencephalogram (EEG). Unique features are defined and extracted from sleep stage data to detect sleep disorders. The entire system consists of the sleep stage detection followed by sleep disorder detection. An accuracy of 98.8% was obtained using an SVM classifier for sleep stage detection and an accuracy of 95.9% for sleep disorder detection using Ensemble bagged tree classifier.