Table 2 FedHNN model performance

	AUC	Accuracy	Precision	Recall	Specificity	F1_Score
All HNN	0.7776	0.7794	0.7974	0.7985	0.7569	0.7961
FedHNN	0.7110	0.7352	0.7319	0.8204	0.6028	0.7598
Single NYU	0.6900	0.7125	0.7689	0.7272	0.6000	0.7696
Single UCLA	0.6880	0.6889	0.6499	0.7917	0.5844	0.7077
Single UM	0.7037	0.7194	0.7272	0.8257	0.5818	0.7664
Single USM	0.6790	0.7010	0.6686	0.6000	0.7583	0.5733
Fed GCN	0.6892	0.7012	0.6881	0.7846	0.5938	0.7192
Fed GAT	0.7033	0.7276	0.7523	0.7468	0.6601	0.7332
Fed GraphSAGE	0.6487	0.6417	0.6605	0.6841	0.6135	0.6536
Fed CNN	0.6893	0.7049	0.7232	0.7244	0.6547	0.7101

Shows the performance of our proposed FedHNN model on four site data. For each strategy and deep learning model, we report the AUC results, accuracy, precision, recall, specificity, and F1 scores for all tasks.

Quick links

Search