A Sequential Model to Recognize Depression Acuity Using Social Media and Physical Activity

Over 350 million people worldwide suffer from depression. A key part in diagnosing depression is screening questionnaires, which rely on patient self-reports of the recent past. With the advent of social media and wearable devices, there is an opportunity for a novel approach to detecting when a patient diagnosed with chronic depression becomes acute. In this project, we use social media data and physical activity data to detect depression acuity. Social media is indicative of an individual’s mental state. Physical activity is an indicator of physical wellness. Our aim is to use machine learning to uncover patterns within and among these two diverse data sources to build a sequential model to identify depression acuity more accurately and sooner than existing screening questionnaires. Dapasoft benefits from this project since it is complementary to their products for handling electronic health records.

Faculty Supervisor:

Rhonda McEwen

Student:

Morteza Zihayat Kermani

Partner:

University of Toronto

Discipline:

Computer science

Sector:

University:

University of Toronto

Program: