Predicting Students' Mental Health Using Fitbit Data

Pins, Noah; Lusignan, Kyle; Escott, Allison; Raisbeck, Olivia

Student Work

Predicting Students' Mental Health Using Fitbit Data

Public Deposited

Mental health screening is critical for early intervention and treatment for anxiety and depression. Rates of depressive or anxious symptoms increased post-COVID-19, particularly in college students, leading to increased difficulties in schoolwork and daily life, and can even lead to more serious conditions. While current screening for mental health may be exhaustive and costly, this study attempts to find an alternate method – using a wearable physiological data watch and taking environmental satisfaction surveys – to be able to create a similar or even more effective and efficient diagnosis process. Utilizing algorithms through machine learning, we were able to see whether this alternate method is usable in its current state, completely unplausible, or simply a viable possibility within reasonable advancements in the years to come. After running multiple tests and performance metrics, this dataset did not explicitly back up the goal to find an alternative screening method, but with further testing and a more encompassing dataset, machine learning techniques could help find connections between physiological and personal data with mental health disorders.

This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

Subject

Publisher