Machine learning for sleep quality analysis

Sleep is an essential and fundamental process of human existence. Low sleep quality or sleep disorders can lead to the worsening of the overall health state, as it plays a crucial role in multiple physical and psychological processes in the human organism. Therefore, understanding health and sleep parameters, that define sleep quality, plays an important role in improving overall health conditions. Methods of machine learning extend the sleep researchers’ toolbox with some novel approaches and possibilities. This paper sums up the main topics, where machine learning has contributed to sleep research in the last decades and describes an example of sleep quality modeling using both physiological and psychological sleep and health characteristics from an open-source dataset. The model is based on the measurements obtained with an actigraph and the potential of alternative devices for data collection is discussed together with the importance of this approach. Model performance and discovered correlations are described in the results section. This overview shows how machine learning techniques can be used to characterize sleep disorders, evaluate the risk factors based on daily measurements and monitor sleep quality.