Automatic Depression Prediction using Screen Lock/Unlock Data on the Smartphone

As COVID-19 continues for a long time, more and more people feel psychologically anxious beyond stuffy. However, as people have high resistance to mental health treatment, there are many cases in which the treatment period with depression is missed and the symptoms are getting worse. In this paper, we study a technology to diagnose users’ depression by using a smartphone that has become an indispensable item carried by the mass of modern people in everyday life. Most of the existing studies diagnosed depression by using the questionnaire responses from smartphone users directly, but this study intends to replace the questionnaire responses from only the mobile phone sensing data without the user’s annoyingness. In particular in this paper, it shows that we can predict the user’s sleep time using only lock/unlock data and detect changes in sleep patterns to predict the likelihood of depression in smartphone users.