Table 2.

The performance results of DR-IBRW and comparing methods on Gottlieb’s dataset [11] and Luo’s dataset [12]

	AUROC	AUPR	Micro-F1	Macro-F1	Precision	Recall
	Gottlieb’s dataset
DR-IBRW	0.955±0.000	0.499±0.174	0.613±0.006	0.513±0.005	0.332±0.002	0.880 ±0.000
MBiRW	0.933 ±0.000	0.213 ±0.028	0.294 ±0.004	0.244 ±0.003	0.256 ±0.001	0.906±0.000
BLM	0.865 ±0.000	0.298 ±0.003	0.583 ±0.001	0.479 ±0.001	0.315 ±0.000	0.891 ±0.000
JI	0.845 ±0.001	0.247 ±0.043	0.385 ±0.003	0.462 ±0.004	0.250 ±0.001	0.894 ±0.181
HGBI	0.811 ±0.000	0.016 ±0.000	0.187 ±0.001	0.157 ±0.001	0.101 ±0.000	0.367 ±0.007
NBI	0.503 ±0.000	0.000 ±0.000	0.022 ±0.000	0.018 ±0.000	0.012 ±0.000	0.039 ±0.001
	Luo’s dataset
DR-IBRW	0.964±0.000	0.529±0.167	0.537±0.006	0.452 ±0.004	0.294±0.002	0.895±0.002
MBiRW	0.945 ±0.000	0.285 ±0.042	0.431 ±0.004	0.363 ±0.003	0.236 ±0.001	0.835 ±0.013
BLM	0.892 ±0.000	0.424 ±0.017	0.527 ±0.003	0.463±0.004	0.278 ±0.001	0.843 ±0.000
JI	0.865 ±0.000	0.287 ±0.041	0.537 ±0.004	0.447 ±0.003	0.294 ±0.001	0.783 ±0.000
HGBI	0.848 ±0.000	0.037 ±0.001	0.170 ±0.001	0.141 ±0.001	0.093 ±0.000	0.318 ±0.005
NBI	0.479 ±0.000	0.000 ±0.000	0.020 ±0.000	0.016 ±0.000	0.011 ±0.000	0.032 ±0.000

The entry in boldface represent the method perform best in this evaluation metric