한국문학 분야 KORMARC 서지레코드를 활용한 저작 식별 개선 방안 연구

초록

This study proposes an improved method for identifying works using KORMARC bibliographic records in the field of Korean literature, provided by the National Library of Korea. Previous research primarily clustered works based on exact matches between bibliographic record sets. However, this approach often led to the separation of identical works due to differences in notation and other variations. To address this issue, this study applied preprocessing techniques that included Chinese character-to-Korean conversion, reordering of Western names (surname first, then given name), and the removal of author role terms and added titles when extracting author names and titles. Additionally, a network analysis technique was employed to construct separate networks based on author names and titles. The process of identifying the intersection of record sets across these networks was then used to identify works and link related records. As a result, 268,684 works were identified from 453,846 bibliographic records, demonstrating a more refined work identification process compared to exact string matching methods. Nevertheless, some limitations were observed in identifying works due to errors in some bibliographic record inputs, highlighting the need for further research to address these issues. The proposed method is expected to improve the work identification process, thereby improving bibliographic entity identification and, in the long term, contributing to the transformation of bibliographic records into linked data.

키워드