New Technology Analyzes Social Media Posts To Detect Suicidal Users

August 24, 2013  |  

With the wave of untimely celebrity deaths, including Glee favorite Cory Monteith, Bachelorette star Gia Allemand, and Disney actor Lee Thompson Young, a much-needed new project has surfaced to help detect suicidal inclinations in social media, reports Mashable.

Called the “Durkheim Project,” it meshes “search technology with predictive analysis to estimate the suicide risk of an individual based on what he or she is posting on Facebook, Twitter or LinkedIn,” Mashable added.

The program will examine users’ social media posts — on Twitter, Facebook, and LinkedIn — to search for red flags that might indicate the user is at risk of self-harm or a mental illness. “[T]he system will also look at other factors that might demonstrate a person’s level of agitation or use of painkillers,” the site said.

“It turns out that what is predictive is much more related to social isolation, chronic pain and things like that — which are actually strongly associated in the literature with suicidal risk,” said Chris Poulin, the project’s principal investigator.

When alarming keywords in status updates begin to correlate with a propensity for suicide or psychiatric illness, an alert for an intervention is sent out. Words like “agitation,” “frightening,” and “Demerol” triggers heightened suspicions that a user may be suicidal.

“Do I think that we can predict suicide risk right now with high precision? No. But I do think that if someone in our system was suicidal today, I think we would have a chance of picking them up,” Poulin said.

Currently, the project is analyzing Facebook users, active duty military members and veterans in particular. Suicides are the second leading cause of death in U.S. service members.

How effective do you think this project will be in preventing suicides?

Trending on MadameNoire

Comment Disclaimer: Comments that contain profane or derogatory language, video links or exceed 200 words will require approval by a moderator before appearing in the comment section. XOXO-MN