10.7910/DVN/649TSE Zong, Chang Chang Zong zhejiang university of science and technology Harvard Dataverse 2026 Engineering Medicine, Health and Life Sciences Evidence Knowledge Graph Information Extraction from Full-Text Literature Biomedical Reasoning Zong, Chang Chang Zong 2026-03-30 2026-03-31 47420159 27687460 2447 127571 80036 72326 73863 128157 838599 23507508 41081656 application/json application/json text/markdown application/json application/json application/json application/json application/json application/json application/json application/json 1.0 Creative Commons CC0 1.0 Universal Public Domain Dedication. This is the public derived data release for EvidenceNet, a framework for constructing disease-specific, evidence-centric knowledge graphs from full-text biomedical literature published during 2010-2025 (around 500 articles for each disease). This package contains two released resources: EvidenceNet-HCC and EvidenceNet-CRC. For each disease, it provides record-level JSON files (evidence_nodes.json) and graph-level JSON files (evidence_graph.json) that preserve structured evidence records, normalized biomedical entities, typed evidence-evidence relations, provenance metadata, and evidence quality scores. The package also includes selected derived evaluation outputs for component validation, question answering, future link prediction, and target prioritization. Raw source PDFs, large third-party knowledge resources, and local cache files are not included.