Sleep and Sedentary Behavior Analysis from Physiological Signals using Machine Learning

In this modern digital world with rapidly changing technologies, personal health, i.e., physical as well as mental health is affected by sedentary behavior. In order to monitor the human sedentary behavior and its analysis requires some knowledge on physical as well as mental disciplines along with some measures of physical exercise as well. The purpose of this paper is to analyze the sedentary, sleep, or the rest behavior of a healthy adult person with the help of the physical activity (PA). The data has been collected from a open data source named as “Replication Data for: Classification of sleep, sedentary behaviour, and physical activity using commercial wearable devices” published on 2019 by Harvard University. The data comparison has been made with various machine learning algorithms. Accuracy has been obtained for each case and the case with the best possible accuracy has been considered for monitoring and analysis purpose. With this kind of analysis it is easy to establish an understandable relationship between parameters that affect the sleep and sedentary behavior with the physiological signals obtained from the commercial wearable devices. The interrelationship will help to scrutinize the hazards from this kind of sedentary behaviour. In such cases, necessary awareness can be build up among the society. The problem can be detected prior to any mishap. Also some detection methods can be applied to reduce such threat and required solution can be suggested as well.