본문 바로가기
  • Home

A Quality Assessment of the National Bibliographic LOD Using SHACL and ShEx: A Dual-Layer Analysis of Ontology Design and Instance Data Validation

  • Journal of the Korean Biblia Society for Library and Information Science
  • 2026, 37(2), pp.321~347
  • DOI : 10.14699//kbiblia.2026.37.2.321
  • Publisher : Journal Of The Korean Biblia Society For Library And Information Science
  • Research Area : Interdisciplinary Studies > Library and Information Science
  • Received : May 21, 2026
  • Accepted : June 4, 2026
  • Published : June 30, 2026

Park Jin Ho ORD ID 1

1한성대학교

Accredited

ABSTRACT

This study aims to diagnose the quality of the National Library of Korea’s national bibliographic Linked Open Data (LOD) at both the ontology design layer and the instance data layer, using SHACL and ShEx, the W3C standard languages for RDF validation. Adopting a complete enumeration approach rather than sampling, the study analyzed the entire dataset of approximately 915 million triples published as of April 1, 2026. The analysis revealed, at the design layer, an incompleteness of specification in which 41.4% of all properties lacked a defined domain, along with the non-incorporation of BIBFRAME vocabulary. At the instance data layer, SHACL validation showed a high conformance rate of 99.73%, yet 99.6% of violations were concentrated in missing titles within the serial (online resources) dataset; ShEx validation showed that the average proportion of unexpected properties between the deductive and inductive shapes reached 98.91%, indicating that the design specification fails to encompass the representational richness of the instance data. In particular, a design-instance asymmetry was identified in which BIBFRAME vocabulary, absent from the design, was extensively used in the instance data.

Citation status

* References for papers published after 2024 are currently being built.