Year: 2026
Type: article
Abstract: The Wikidata ontology consists of classes, properties, and subproperties which provide structure to an open and flexible graph-based semantic network. It includes nearly every class in Wikidata, resulting in a large, dynamic ontology. Sourced from concepts written about in Wikipedias across a wide variety of languages, Wikidata is arguably the most reliable openly licensed and actively edited knowledge graph in the world, with growing interest in Wikidata reuse by applications like the Blue Core Graph Toolbox, using Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) and a vector database created from a sample Library of Congress Bibliographic Framework (BIBFRAME) data to produce more relevant responses for librarians creating and editing Resource Description Framework (RDF) descriptions for library materials. This novel use case highlights the need for more consistent modeling across the Wikidata ontology, and within specific domains of knowledge. The bibliographic domain serves as a case study for surveying the types of ontology issues present in Wikidata. Potential solutions, such as aligning the Wikidata bibliographic domain with the IFLA Library Reference Model (LRM) create pathways to building community consensus and increasing the likelihood of Wikidata reuse within the Galleries, Libraries, Archives, and Museums (GLAM) community.
Source: Journal of Open Humanities Data
DOI: 10.5334/johd.454
Institutions: Stanford University, University of Würzburg
Language: English
Topic: Semantic Web and Ontologies
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $367
Funder: Wikimedia Foundation
OpenAlex ID: W7118925302