University of Bristol News (03/04/15)
A paper by researchers at the University of Bristol describes their use of big data analytics to understand and characterize the media coverage of the 2012 U.S. presidential election. Academics from the university's Intelligent Systems Laboratory (ISL) analyzed more than 130,000 online news articles using a big data system featuring new methods that are richer than traditional word-association networks. The researchers used a semantic graph that linked a given text to identified noun phrases and verbs, using the resulting subject-verb-object triplets as the basis for a word-association network. The researchers note this method has never been used with such a large real-world dataset, and employing it enabled them to show, for example, that Democrats presented a more united front on policy issues, while Republicans were divided on several issues. "Mapping the full electoral campaign coverage by offline and online media is a very difficult challenge, given the large amount of data and the large number of sources available in advanced democracies," says ISL's Saatviga Sudhahar. "We believe that the methodology used for the study is a big step forward in the linguistic analysis of texts by using extracted relational data and could help us understand major events."
No comments:
Post a Comment