도서관목록의 주제 접근성 향상을 위한 주제명 데이터의 품질 비교

The purpose of this study was to make a comparative analysis of the quality of subject data constructed by library, bookstore, social tagging site. For this purpose, a comparison was made between subject headings by the National Library of Korea and Library of Congress(LC) and subjects or tags by Kyobo bookstore and LibraryThing. To the end, 24 self-development books, 24 novels, and 24 history books were randomly selected; then, subject headings, subjects, and user tags were drawn from each web site to make quantitative analysis, pattern analysis, and content analysis. The findings from these analyses were as follows: First, the quantitative analysis found that LibraryThing(197.8 on average) had the largest number of subject data, followed by Kyobo bookstore(19.2), LC(3.9), and the National Library of Korea(1.28). Second, the pattern analysis found that subject data constructed by the National Library of Korea belonged to a single pattern in most cases. In contrast, the other sites commonly had various patterns as topic, localities/countries, genre, and person, even enough there were different rates among subject areas. Third, the content analysis found that common data had different characteristics by selected works. The self-development books had terms indicating the persons or the characters mentioned in the work and those indicating lessons or the author’s intentions. Novels had terms indicating the genre and key words indicating the contents of the works. The history books had a term ’history’ and terms indicating places or countries in the work. The result of this study will be used for the development of guideline for constructing subject data in Korea libraries.