RE-SAMPLE provides open-access synthetic data, generated from the real datasets collected in the project, that have similar statistical distributions to the real data. Furthermore, RE-SAMPLE makes available as open access models trained on that syntehtic data. Both are published in the RE-SAMPLE community on Zenodo.
You can explore and download available datasets, as well as view schemas and instructions. The direct link to the dataset is DOI:10.5281/zenodo.17106631.
- To ensure interoperability, synthetic datasets align with the schema of the Health Data Hub (HDH) in the Edge Nodes (also published open-access in the RE-SAMPLE community on Zenodo).
- The data follow the FHIR (Fast Healthcare Interoperability Resources) V4.3.0 standard and are available in CSV, and the schema is available as a JSON.
You can explore and download available pre-trained machine learning models trained on synthetic datasets, as well as documentation and integration guide. The direct link to the dataset is DOI:10.5281/zenodo.17106934.
- The models are available in standard formats compatible with widely used machine learning frameworks, such as the Flower Framework and PyTorch, ensuring easy adoption by researchers and practitioners.