Company researchers create new technology to predict disasters

Feb 5, 2013 12:23 GMT  ·  By

Microsoft and researchers at the Technion – Israel Institute of Technology have developed a new technology that can predict disasters based on old news reports.

The new software uses New York Times articles and online data provided by several other sources to issue warning regarding disease outbreaks, natural disasters and violence, according to a report published by MIT Technology Review.

In addition to a 22-year New York Times article database, the new technology also uses data from DBpedia, WordNet and OpenCyc. There are more than 90 sources included at this point, Eric Horvitz, one of the Microsoft researchers involved in the project, revealed.

“One source we found useful was DBpedia, which is a structured form of the information inside Wikipedia constructed using crowdsourcing. We can understand, or see, the location of the places in the news articles, how much money people earn there, and even information about politics,” Kira Radinsky, a PhD researcher at the Technion-Israel Institute, added.

A greatly improved version of the system could be used in real life, Horvitz explained, most likely to assist experts in predicting disasters, but also to help humanitarian response.

“We’ve done some reaching out and plan to do some follow-up work with such people,” Horvitz told the source.

As far as testing sessions go, the newly developed system provides impressive results. A cholera warning was issued when it analyzed reports of droughts in Angola in 2006 and storms in Africa in early 2007. In both cases, cholera outbreaks were confirmed by the media approximately one week later.

According to researchers, the technology is right in 70 to 90 percent of the cases.

Just as expected, the Redmond-based technology giant has no plan to sell the newly developed technology, as researchers only intend to help various government agencies and organizations around the world in planning humanitarian response to the predicted disasters.