We typically think of predictive analytics being used for providing business insight into opportunities such as upselling customers or detecting risk of fraud, but Patterns and Predictions, a predictive analytics company, is putting its technology to work for a higher purpose. The company is working on a big data initiative dubbed “The Durkheim Project” which is using its machine learning data fabric built on technologies from Attivio and Cloudera to predict suicide at scale.
Patterns and Predictions is now providing technology enabling opt-in participation from more than 100,000 U.S. veterans to build a ‘big’ medical database that will help military mental health experts combat the incidence of suicide among veterans. Cited by a 2012 TIME magazine cover story as a problem of epidemic proportions, suicide rates among veterans are roughly double those of adults in the general U.S. population.
Patterns and Predictions’ founder Chris Poulin, has been working with Dartmouth researchers to address this problem since 2010. “The next generation of predictive analytics tools gives scientific and clinical investigators new hope and resources for solving even the most intractable problems,” says Poulin.
In 2011, Patterns and Predictions engaged big data analytics experts Attivio and Cloudera and secured a contract with the Defense Advanced Research Projects Agency (DARPA). Concluded in February 2013, staff from Patterns and Predictions, Dartmouth, and the U.S. Veterans Administration conducted an investigation to validate the machine learning data fabric. Initial findings indicated the data model’s predictive accuracy was statistically significant (consistent accuracies of 65% or more) in predicting suicidality risk among a veteran control group.
Having confirmed the theoretical underpinnings of this work, Patterns and Predictions is leading “Phase 2” technology for The Durkheim Project, a big data initiative with only one objective: suicidality prediction at scale. “The promise of Durkheim lies in its ability to collect and monitor a diverse repository of complex data, with the hope of eventually providing a real-time triage of interventional actions upon detection of a critical event,” says Poulin.