This Software Can Analyze News And Predict Disease, Violence, And Disaster

Flickr/caphotosnewyork

Researchers from Microsoft and the Technion-Israel Institute of Technology have developed software that can predict when and where outbreaks might occur, Tom Simonite of MIT Technology Review reports.

The system, which taps into 22 years worth of archives from The New York Times, works by identifying and analyzing clues from news reports and more than 90 other data sources from the web.

The prototype software can predict things like disease outbreaks, violence, and natural disasters.

For example, when tested on historical data from 2006, the system correctly predicted a cholera outbreak based on reports of droughts in Angola. The software is right between 70 to 90 percent of the time.

Eric Horvitz, a scientist and co-director at Microsoft Research, told Simonite that a refined version of the system could be used in real settings to help experts prepare for emergency and humanitarian response, for example.