Predictive Analytics for Mental Health Outcomes: A Case for Novel Applications of the Hedonometer
A HIMSS18 presentation
Grounded in behavioral experimentation, the Hedonometer is an instrument to remotely-sense emotional states and levels, in real time or post hoc. While well suited for the analysis of social media, the Hedonometer provides an analytical platform to perform large-scale, sentiment analysis on many corpora in 10 global languages. The development of the Hedonometer arose from a longstanding collaboration between the Computational StoryLab at the University of Vermont (UVM) and The MITRE Corporation, a collaboration which also produced the hedonometer.org web service. Since the deployment of the web service, the UVM team has explored various applications of the Hedonometer as an initial screening tool for detecting mental health outcomes in advance of clinical diagnoses. Here, we will detail our latest collaborative pursuit: applying the Hedonometer as an initial screening tool to identify those veterans at greatest risk to attempt suicide.
Speakers: Josh Park, MITRE and Lisa Tompkins-Brown, MITRE