A Novel Curated Scholarly Graph Connecting Textual and Data Publications

• DOI: 10.1145/3597310
• Publication Date: 2023-05-19T09:49:00Z
• Authors: Ornella Irrera; Andrea Mannocci; Paolo Manghi; Gianmaria Silvello
• Keywords: Linked Data; Knowledge graph; Scholarly Communication

In the last decade, scholarly graphs became fundamental to storing and managing knowledge in a structured machine-readable way. Methods tools for discovery impact assessment of science rely on such their quality serve scientists, policymakers, publishers. Since research data very important communication, started including dataset metadata relationships publications. Such are foundations Open Science investigations, data-article publishing workflows, discovery, indicators. However, due heterogeneity practices (FAIRness is indeed making), they often lack complete reliable necessary perform accurate analysis; e.g., inaccurate, author names not uniform, semantics unknown, ambiguous or incomplete. This work describes an open curated graph we built published as training test set connection, disambiguation, link prediction tasks. Overall contains 4,047 publications, 5,488 datasets, 22 software, 21,561 authors; 9,692 edges interconnect publications datasets software labeled with that outline whether publication citing, referencing, documenting , supplementing another product. To ensure high-quality semantics, relied information extracted from PDFs webpages curate enrich nodes semantics. best our knowledge, this first ever resource, manually validated metadata.