Node Importance Estimation with Multiview Contrastive Representation Learning

Table 3

Experimental results on real-world datasets.

Method	FB15K			TMDB5K			IMDB
	NDCG @100	SPEARMAN	HR @100	NDCG @100	SPEARMAN	HR @100	NDCG @100	SPEARMAN	HR @100
PR	0.8395 0.009	0.3505 0.018	0.1340 0.015	0.8378 0.016	0.6092 0.012	0.4140 0.064	0.8485 0.032	0.1829 0.003	0.3720 0.017
PPR	0.8407 0.009	0.3656 0.022	0.1360 0.019	0.8594 0.007	0.7248 0.013	0.4220 0.060	0.8638 0.029	0.3979 0.004	0.4060 0.019
LR	0.8921 0.011	0.6097 0.017	0.2160 0.042	0.8443 0.015	0.6871 0.011	0.4300 0.034	0.8972 0.003	0.5838 0.003	0.4540 0.024
RF	0.9136 0.010	0.6441 0.009	0.2220 0.048	0.8617 0.014	0.6993 0.016	0.4680 0.028	0.9068 0.006	0.6112 0.004	0.4380 0.018
GCN	0.9408 0.007	0.7179 0.013	0.4240 0.051	0.8999 0.008	0.7736 0.011	0.5240 0.029	0.9031 0.009	0.7135 0.006	0.3240 0.035
GAT	0.9357 0.014	0.6898 0.047	0.4060 0.055	0.9017 0.010	0.7728 0.008	0.5360 0.031	0.9186 0.008	0.6942 0.007	0.4780 0.028
GENI^∗	0.9301 0.005	0.7385 0.012	0.4260 0.092	0.9018 0.006	0.7898 0.009	0.5460 0.024	0.9302 0.005	0.7312 0.005	0.5080 0.034
GENI	0.9415 0.006	0.7797 0.018	0.4220 0.058	0.8979 0.008	0.7721 0.007	0.5340 0.021	0.9394 0.005	0.7351 0.005	0.5420 0.026
RGTN	0.9501 ± 0.007	0.8156 ± 0.010	0.4880 ± 0.064	0.9114 ± 0.009	0.7946 ± 0.010	0.5580 ± 0.036	0.9585 ± 0.004	0.7643 ± 0.003	0.5660 ± 0.038
HIVEN	0.9436 0.006	0.7673 0.015	0.4280 0.052	0.9045 0.008	0.7852 0.008	0.5480 0.026	0.9388 0.006	0.7475 0.006	0.5340 0.036
MCRL	0.9626 0.013	0.8229 0.014	0.5100 0.069	0.9195 0.020	0.7961 0.013	0.5680 0.041	0.9624 0.008	0.7713 0.007	0.5720 0.044

The bolded results are the best, and the italic results are the second best.