Data-Driven Support for Substance Addiction Recovery Communities

摘要

This project aims to assist those suffering from substance addiction so that they might better understand their mental health. I propose a user level in-app relapse report classifier that has an overall accuracy of 0.882 in a dataset where 18% of users report a substance relapse. Interactions between engagement and views of inspirational messages of the day are complex, yet significant. Predictive analytics inspire design implications for sociotechnical systems, which aim to facilitate the addiction recovery process.