BMC Medical Informatics and Decision Making

Table 2 Classification effectiveness with balanced sample nodes

From: Health insurance fraud detection by using an attributed heterogeneous information network with a hierarchical attention mechanism

Train:Val:Test	Metrics	Metapath2Vec	GAT	GCN	HAN	MHAMFD_hierar	MHAMFD
1:1:3	f1	0.5595	0.7979	0.7914	0.791	0.8197	0.8694
1:1:3	acc	0.7196	0.8523	0.8523	0.8598	0.8611	0.8961
2:1:2	f1	0.5964	0.7273	0.7547	0.7455	0.8354	0.8566
2:1:2	acc	0.7386	0.8295	0.8523	0.8409	0.8678	0.8905
3:1:1	f1	0.6349	0.72	0.7772	0.7407	0.8123	0.8439
3:1:1	acc	0.7386	0.8409	0.8371	0.8409	0.8423	0.8757

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com