TABLE 3.

Performance compared with 7 baseline methods on TL-HGBI dataset.

Methods	AURP	AUC	F1	Acc	Rec	Spe	Pre
BNNR	0.4502 ± 0.002	0.9065 ± 0.001	0.4640 ± 0.003	0.9462 ± 0.000	0.5098 ± 0.020	0.9671 ± 0.015	0.4261 ± 0.004
DRHGCN	0.4824 ± 0.000	0.9295 ± 0.032	0.4723 ± 0.016	0.9442 ± 0.035	0.5459 ± 0.004	0.9633 ± 0.000	0.4161 ± 0.000
DRWBNCF	0.3432 ± 0.001	0.8927 ± 0.004	0.4013 ± 0.000	0.9306 ± 0.002	0.5090 ± 0.002	0.9508 ± 0.030	0.3314 ± 0.001
LAGCN	0.4970 ± 0.004	0.9155 ± 0.005	0.4586 ± 0.007	0.9413 ± 0.002	0.5440 ± 0.033	0.9603 ± 0.000	0.3968 ± 0.001
NIMCGCN	0.1532 ± 0.003	0.7490 ± 0.018	0.2317 ± 0.020	0.9012 ± 0.000	0.3265 ± 0.005	0.9287 ± 0.001	0.1802 ± 0.006
DDA-SKF	0.2266 ± 0.015	0.8608 ± 0.007	0.3136 ± 0.000	0.9071 ± 0.001	0.4646 ± 0.003	0.9283 ± 0.002	0.2368 ± 0.003
REDDA	0.4243 ± 0.000	0.9225 ± 0.001	0.4493 ± 0.013	0.9406 ± 0.001	0.5308 ± 0.002	0.9602 ± 0.025	0.3898 ± 0.010
LBMFF	0.5261 ± 0.003	0.9160 ± 0.000	0.4821 ± 0.002	0.9429 ± 0.003	0.5898 ± 0.002	0.9596 ± 0.005	0.4078 ± 0.007

The best results are in bold faces and the second-best results are underlined.