Technical Report: Curation Modeling Guideline for the Open Research Knowledge Graph

Abstract

The Open Research Knowledge Graph (ORKG) enables the structured representation and reuse of scholarly knowledge across scientific domains. As the platform grows, maintaining consistent, interoperable, and semantically meaningful knowledge representations becomes essential. This report presents the first iteration of the ORKG modeling guideline, a set of practical recommendations designed to support contributors and curators in creating high-quality knowledge graph content. The guideline addresses common modeling challenges, including duplicate entities, inconsistent naming, incomplete metadata, and unstructured representations. It introduces 13 recommendations organized into four categories: Reuse Quality, Content Quality, Modeling Quality, and Representation Quality. Each recommendation provides practical guidance on how to apply the principle, its expected impact, and the rationale behind it. Developed collaboratively by the ORKG Curation and Community Building (C&CB) team, the guideline serves as a living resource that will evolve with emerging modeling needs and community experience. By promoting consistent and reusable modeling practices, it aims to improve the quality, sustainability, and usability of scholarly knowledge representations in the ORKG.