FedGraphHE experimental data and preprocessed graph features

This repository contains the processed datasets used in the experiments of the manuscript “FedGraphHE: A privacy-preserving federated graph neural network framework with dynamic homomorphic encryption and robust aggregation”.The archive includes preprocessed versions of the Cora, CiteSeer and PubMed citation network benchmarks, as well as the derived data used for our ISIC 2020 dermoscopic image experiments. For the citation networks, we provide commonly used processed graph files and tensors that can be loaded directly for training and evaluation. For the ISIC 2020 setting, we share the image identifiers, labels, train/validation/test splits, feature representations and the graph structure constructed from these features.All underlying raw datasets (the ISIC 2020 Challenge data and the original Cora, CiteSeer and PubMed benchmarks) are publicly available from their official providers. The data in this repository are the corresponding processed versions used in our work and are intended to facilitate replication and reuse of the reported results.