Towards Generating Synthetic EHR Knowledge Graphs — a Probabilistic Approach
Date Issued
2025-06-12
Author(s)
Milenkova, Eva
Jakubowski, Maxime
Hose, Katja
Abstract
Advances in medical AI and data analytics require large amounts of patient data. Due to privacy concerns, such data is not always available. Synthetic data generation promises a solution to provide the required data despite privacy restrictions. In this paper, we therefore introduce SynMedRDF, an open-source tool to generate synthetic Electronic Health Records. It ensures clinical accuracy by using real-world probabilities and correlations. The data is output as an RDF knowledge graph, enabling structure- and semantics-aware sharing, linking, and analysis.
