FedEdge-6G Synthetic Federated Learning Dataset (v1.0)

This repository contains the synthetic dataset used in our paper “FedEdge-6G: Cost-Optimised, FLTrust-Hardened Federated Learning for 6G Edge Networks.” The data emulate latency-constrained, non-IID client partitions typical of vehicular and smart-factory environments. <br><br> Contents <ul><li>/clients/ – 100 JSON files, one per simulated client, each with 10 000 labelled samples (features ≈ channel state + task demand; target ≈ regression value).</li><li>README.md – schema and generation script parameters.</li><li>gen_config.yaml – reproducible seed & distribution settings.</li></ul> <br> Intended use <br>Benchmarking federated-learning algorithms under non-IID, bandwidth-limited, and adversarial settings. <br> Ethics / limitations <br>No real user data; entirely synthetic. Not suitable for evaluating privacy leakage on real-world sensitive attributes.