Neighborhood Attentional Memory Networks for Recommendation Systems

Table 3

Experimental results for different methods on the Epinions, citeulike-a, and Pinterest datasets.

	Epinions				Citeulike-a				Pinterest
	HR@5	HR@10	NDCG@5	NDCG@10	HR@5	HR@10	NDCG@5	NDCG@10	HR@5	HR@10	NDCG@5	NDCG@10
BPR	0.5584	0.6659	0.4334	0.4683	0.6547	0.8083	0.4858	0.5357	0.6936	0.8674	0.4912	0.4912
GMF	0.5365	0.6562	0.4016	0.4404	0.7271	0.8326	0.5689	0.6034	0.6726	0.8505	0.4737	0.5316
FISM	0.5542	0.6717	0.4192	0.4573	0.6727	0.8072	0.5106	0.5545	0.6783	0.8654	0.4658	0.5268
KNN	0.1549	0.1555	0.1433	0.1435	0.6990	0.7348	0.5789	0.5909	0.5738	0.8376	0.3450	0.4310
SVD++	0.5628	0.6754	0.4112	0.4477	0.6952	0.8199	0.5244	0.5649	0.6951	0.8684	0.4796	0.5362
NeuMF	0.5500	0.6660	0.4214	0.4590	0.7629	0.8647	0.5985	0.6316	0.7041	0.8732	0.4978	0.5530
CMN-2	0.6017	0.7007	0.4724	0.5045	0.7959	0.8921	0.6185	0.6500	0.7267	0.8904	0.5180	0.5714
NAMN-2	0.6230	0.7122	0.4913	0.5234	0.8233	0.9101	0.6576	0.6867	0.7470	0.8966	0.5432	0.5918