Relating Keywords to the Top Ten News of the Year in Korean Newspapers

PACLIC 28 !28 Relating Keywords to the ‘Top Ten News of the Year’ in Korean Newspapers Jae-Woong Choe Korea University / Seoul, Korea jchoe@korea.ac.kr Abstract This paper takes an in-depth look at the relationship between mechanically extracted keywords and ‘Top Ten News of the Year’ compiled by the news editors. A previous study that briefly touched on the topic concludes there does not seem to exist any meaningful connection between the two. In this paper, we set up a more elaborate way of comparing and connecting the two, and argue that there is a certain reasonably good converging point. The corpus we make use of for our experiment is a subset of the Trend 21 corpus which is a collection of Korean major newspapers (2000-2013). For keyword extraction, log- likelihood ratio was made use of. Extraction of collocation for each keyword was needed, for which a version of Mutual Information was utilized. Finally a detailed comparison of the top ten news with the top 100 keywords was conducted from several points of view. 1 Introduction There is a growing use of the keyword methodology as an analytic tool to efficiently analyze texts or corpora, and we can say it now has established itself as a viable, and, importantly, objective alternative to the traditional and rather introspective method of discourse or cultural interpretation (Scott & Tribble, 2006, Bondi & Scott, 2010; Archer, 2009; Baker et al., 2013). One issue that needs to be addressed is how the new methodology relates to the introspective way of selecting keywords from a given text or corpus. Are the two supposed to be different from each other? Then why so? Or do they have to be comparable or even identical to each other? And if they do, how can we test the comparability or convergence? In this paper we raise these questions on the basis of Korean newspaper corpora and some lists of ‘Top Ten News (T10N for short)’ compiled, presumably introspectively, by the newspaper editors at the end of every year. Specifically we compare the top 100 keywords with T10N, and see how well they converge with each other. Some previous studies seem to suggest a tentatively negative conclusion on the issue of any systematic relationship between the introspective and quantitative keywords in general (Bondi & Scott, 2010), or between quantitative keywords and T10N (Kim & Lee, 2011). It should be noted at the outset that each item in T10N is not a keyword per se. Though it can trivially turned into a small set of keywords on the basis of the surface description of the news item, we take the item as an abstract concept to which a set of keywords can be mapped, thus making it possible to compare the quantitatively derived keywords and the more abstract key concepts that are selected introspectively. We argue that T10N provides an interesting testing ground for the significance of the keywords extracted or the role of the ‘human factor’ in the selection process of T10N. …

Content truncated to 500 words