Identifying associations between smartphone usage and mental health during depression, anxiety and stress

Reality mining in behavioral studies is one of the highly researched topics in computational science. The reason behind this great deal of experimentation is the particularly dynamic and potentially latent nature of human behavior. In this work, we have showcased our efforts in analyzing the associations between the smartphone usage and the mental state of the person during depression, stress and anxiety. We have particularly focused on descriptive modeling of data using multiple linear regression to find correlations between the passive smartphone data stream and the user filled survey. Finally we estimate the predictive capability and outline the major challenges of such a system.