Applied Network Science

Table 2 Hits@1

From: Relation prediction in knowledge graph by Multi-Label Deep Neural Network

	hits@1
Method	FB15k		WN18		WD40k		WD40k_nl
	raw	filt	raw	filt	raw	filt	raw	filt
TransE	0.651	0.843	0.586	0.586	0.201	0.205	-	-
TransR	0.421	0.502	0.713	0.714	0.0951	0.0981	-	-
PTransE	0.695	0.936	-	-	0.858	0.890	0.863	0.893
DKRL	0.685	0.872	-	-	-	-	-	-
KGML (α=1.0)	0.725	0.943	0.975	0.980	0.873	0.907	0.907	0.908
KGML (α=0.5)	0.725	0.933	0.972	0.977	0.877	0.911	0.910	0.911
KGML (α=0.0)	0.667	0.963	0.975	0.979	0.879	0.914	0.914	0.916

Underlining in the table means the best accuracy in each method

Back to article page