FedGenCDSS: Federated GenAI Dataset for Clinical Decision Support

FedGenCDSS is a simulated dataset designed for federated Generative AI research in clinical decision support. It contains over 10,000 synthetic Electronic Health Record (EHR) entries split into training and test sets, along with more than 1,000 structured samples in JSON. Each record simulates clinical notes, diagnoses, and metadata across decentralized institutions. The dataset supports benchmarking of privacy-preserving transformer-based models for summarization, classification, and explainable diagnostics. It also includes evaluation scripts, accuracy plots, and benchmarking figures for reproducibility in IEEE healthcare AI research.