Faculty Advisor

Rundensteiner, Elke A.

Abstract

Our project goal was to develop a depression sensing application that leverages multi-modal data sources collected from a smartphone, focusing on features extracted from audio, text messages, social media data, as well as GPS modalities. We conducted extensive experiments to study the effectiveness of these features to improve our machine learning model. We deployed our EMU app on Amazon Mechanical Turk for crowd-sourced data collection and incorporated feature extraction techniques and machine learning algorithms to reliably predict levels of depression.

Publisher

Worcester Polytechnic Institute

Date Accepted

March 2019

Major

Computer Science

Project Type

Major Qualifying Project

Accessibility

Unrestricted

Advisor Department

Computer Science

Share

COinS