A new supercomputer can now forecast global revolts by analysing national moods. Researchers fed a ‘super’ PC articles which then accurately predicted uprisings based on the information.
Click to enlarge
Millions of news articles used in the experiment – 100 million in total – charting national sentiment prior to the Arab Spring revolts in both Egypt and Libya, were fed to a central system and predicted trouble in both countries, could now be used to predict future global events.
How? By analysing media and other reports. The graphed results showed a notable drop in national mood before the violent outbreaks that resulted in Arab Spring uprising earlier this year.
Words like ‘nice’ or ‘terrible’ were searched on a range of articles from BBC Monitoring service to the New York Times and then linked to geocodes and plotted on map coordinates, which helped pinpoint trouble spots.
The analysis, although carried out on ‘supercomputer’ SGI Altix, known as ‘Nautilus’ after the events occurred could prove a important tool to identify future unrest, researchers believe.
Scientist Kalev Leetaru, from the University of Illinois’ Institute for Computing in the Humanities, Arts and Social Science who carried out the experiment, says the same processes could be used to anticipate upcoming conflict and other major global events.
Leetaru, who presented his findings in the journal First Monday, compared the study to weather forcasting. “I liken it to weather forecasting. It’s never perfect, but we do better than random guessing.”
“It is very similar to what economic forecasting algorithms do.”
Nautilus also predicted the location of on-the-run terrorist Osama Bin Ladin within 200km of where he was actually found.
The formidable SGI Altix runs off 1024 Intel Nehalem cores have a total processing power of 8.2 teraflops, according to BBC.
But its not just news articles that are being used to examine global sentiment. Stock market analysts like TweetTrader.net are using Twitter to predict future movements on share markets by analyzing daily tweets posted by millions of users to measure overall sentiment.