Scientists piece together EU media structure
How influential is the media in shaping the news agenda in EU Member States? Very, say EU-funded researchers, who evaluated more than 1 million news articles in 22 languages to identify the factors that make this impact felt. The study, the first mega content analysis of cross-linguistic text using artificial intelligence tools, is presented in the journal PLoS ONE. The research was funded in part by the PASCAL2 ('Pattern analysis, statistical modelling and computational learning 2') project, which received EUR 6 million under the 'Information and communication technologies' (ICT) Theme of the Seventh Framework Programme (FP7). In addition, the software used in the study was developed by the SMART ('Statistical multilingual analysis for retrieval and translation') project. SMART clinched EUR 2.34 million under the 'Information society technologies' (IST) Thematic area of EU's Sixth Framework Programme (FP6) and created a statistical machine translation pipeline for EU news - called 'Found in Translation' - that could demonstrate machine translation technology. Europe-based news outlets cover myriad stories chosen from an extensive repository every day. One may assume that the outlets base their selection on specific criteria, but the researchers found that patterns materialise when all of these choices are assessed over both an extended period and a set of outlets. The computer scientists, led by Professor Nello Cristianini of the University of Bristol in the UK, in collaboration with Professor Justin Lewis of Cardiff University in the UK, found that the news content selected by outlets reflects national bias, and cultural, geographic and economic links between nations. A case in point, say the researchers, is outlets swapping information with one another about common interests, whether it is news about the euro area or who is singing what in the annual Eurovision song contest. They point out that deviation from 'normal content' is evident in outlets of countries that do not share the euro currency, for instance. Research on this topic was relatively non-existent in the past because no tools had been developed to evaluate the vast amount of data available. But thanks to the creation of machine translation and text analysis, the scientists used automated methods from artificial intelligence in their study with fruitful results. 'Automating the analysis of news content could have significant applications, due to the central role played by the news media in providing the information that people use to make sense of the world,' explains Professor Cristianini from Bristol's Intelligent Systems Laboratory. For their study, the researchers pulled articles from the online news feeds of the top 10 news outlets (by volume of web traffic) for each EU Member State. They collected 1,370,874 news items in total over a 6-month period (between August 2009 and January 2010). The number of non-English language news items amounted to 1.2 million; these were translated automatically into English. The team found various links between Greece-Cyprus; Czech Republic-Slovakia; Latvia-Estonia, Belgium-France; and Ireland-UK. 'This approach has the potential to revolutionise the way we understand our media and information systems,' Professor Lewis points out. 'It opens up the possibility of analysing the mediasphere on a global scale, using huge samples that traditional analytical techniques simply couldn't countenance. It also allows us to use automated means to identify clusters and patterns of content, allowing us to reach a new level of objectivity in our analysis.' Professor Cristianini says the tool developed by the SMART project is now used as part of the main pipeline that assesses EU news content. The Institute for the Protection and Security of the Citizen (IPSC), one of the seven institutes of the European Commission's Joint Research Centre (JRC), made a strong contribution to this study.
Countries
United Kingdom