Applying Deep Learning for Prediction Sleep Quality from Wearable Data

Abstract

Sleep is not only very important for physical health but also the mental health of human, that was addressed by many previous studies. Today, with the development of technology, which opens in the application for improving quality of sleep, such as wearable devices, artificial intelligence, neural network. In this study, we applied deep learning (DL) neural networks and smart wearable devices to predict the quality of sleep. The data was collected on students (mean age = 20.79) during 106 days by Fitbit Charge HR™ device. The results showed DL models could predict sleep quality base on physical activities in awake time.