Our Project

Current State of Art:

Diagnosis of psychological health and the prediction of negative events, such as suicide, or their ideation are limited by:

  • A lack of broad societal attention to problem & clearly defined signals for interventions.
  • A lack of real time, near real time reaction capability to large volumes of data.

The Need:

Broader coverage of suicide risk detection and better understanding of the expression of suicide ideation through data mining of text and images.

Proposed Solution:

Continuous monitoring of social network user behavioral intent enabling intervention, facilitated by social/online/mobile data sources.

Research Accomplishments:

  • ‘Behavioral Risk’ clinical dashboard POC (June 2012)
  • Mobile Opt In Framework POC (June 2012)
  • IRB Approval for VA Study (Dec 2012)
  • VA data set analyzed, with predictive models built that have initial accuracies of between 60 – 65% with Features (keywords) extracted, in line with medical literature (Jan 2013)
  • Scaling efforts/outreach with Facebook (Mar 2013)

Current Efforts:

  • Outreach to Active Duty and Veteran Populations for Data Collection efforts
  • ‘Narrative of Suicide’ Research
  • Intervention Protocols Research
View Demo