An adversarial learning approach for discovering social relations in human-centered information networks

EURASIP Journal on Wireless Communications and Networking

Table 4 Micro- F₁ and Macro- F₁ scores on HepTh dataset in link prediction (α=0.9,β=0.1)

%Training edges	10%		30%		50%		70%		90%
	Micro- F₁	Macro- F₁	Micro- F₁	Macro- F₁	Micro- F₁	Macro- F₁	Micro- F₁	Macro- F₁	Micro- F₁	Macro- F₁
GAN	0.7590	0.7553	0.7541	0.7507	0.7472	0.7439	0.7555	0.7525	0.7535	0.7497
LSGAN	0.7714	0.7692	0.7686	0.7666	0.7728	0.7711	0.7452	0.7433	0.7562	0.7546
DeepWalk	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788
node2vec	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788	0.6098	0.3788
AIDW	0.6098	0.3805	0.6091	0.3785	0.6105	0.3840	0.6098	0.3837	0.6105	0.3840
Jaccard	0.6126	0.3865	0.6126	0.3865	0.6126	0.3865	0.6126	0.3865	0.6126	0.3865
CosSim	0.3909	0.2847	0.3909	0.2847	0.3909	0.2847	0.3909	0.2847	0.3909	0.2847
SRGAN	0.7887	0.7844	0.7901	0.7861	0.8004	0.7973	0.7859	0.7820	0.7887	0.7861