September 12, 2011
Can Supercomputers Predict World Events?
By using a supercomputer to analyze news stories, a University of Illinois researcher believes that he can predict major upcoming global events, according to various news and wire reports.
Writing in the journal First Monday on September 5, Kalev Leetaru, the Assistant Director for Text and Digital Media Analytics at the school's Institute for Computing in the Humanities, Arts, and Social Science, writes, "Pooling together the global tone of all news mentions of a country over time appears to accurately forecast its near—term stability, including predicting the revolutions in Egypt, Tunisia, and Libya, conflict in Serbia, and the stability of Saudi Arabia."
Using a shared-memory supercomputer known as Nautilus, Leetaru entered a total of 100 million articles, originating from sources all around the world and first compiled in one of three different news archives. Included in those databases were complete issues of the New York Times from 1945 through 2005, the unclassified edition of Summary of World Broadcasts from 1979 through last year, and an English-language archive of Google News stories from the past five years.
"Advanced tonal, geographic, and network analysis methods were used to produce a network 2.4 petabytes in size containing more than 10 billion people, places, things, and activities connected by over 100 trillion relationships, capturing a cross-section of Earth from the news media," a September 7 press release describing the findings said.
"A subset of findings from this analysis were then reproduced for this study using more traditional methods and smaller-scale workflows that offer a model for a new class of digital humanities research that explores how the world views itself," it added.
According to a Friday BBC News article, Nautilus would have "charted deteriorating national sentiment ahead of the recent revolutions in Libya and Egypt," as well as detect "early clues about Osama Bin Laden's location." The British news service notes that, "while the analysis was carried out retrospectively, scientists say the same processes could be used to anticipate upcoming conflict."
In his paper, "Culturomics 2.0: Forecasting Large-Scale Human Behavior Using Global News Media Tone in Time and Space," Leetaru confirmed that the computer would have forecasted revolutions in both Egypt and Libya, as well as in Tunsia.
Likewise, he notes that it would have correctly identified Bin Laden's projected whereabouts to a 200km radius in Northern Pakistan that would have included Abbotabad, the city where the Taliban leader was killed by U.S. Navy Seals. Leetaru also claims that the computer verifies commonly held assertions that news coverage is becoming more negative, and the American news reports tends to portray a self-centered, nationalistic view of world events.
Image Caption: Global geocoded tone of all Summary of World Broadcasts content, January 1979—April 2011 mentioning “bin Laden”.
On the Net: