A Knowledge Graph Entity Disambiguation Method Based on Entity-Relationship Embedding and Graph Structure Embedding

Table 2

F1 mean values of EDEGE and the baseline method on different data sets.

Method	AIDA-B	AIDA-CoNLL	MSNBC	AQUAINT	ACE2004	Average
PBoH	0.752	0.741	0.753	0.761	0.749	0.7512
DoSeR	0.895	0.885	0.725	0.876	0.883	0.8528
NeuPL	0.896	0.878	0.879	0.884	0.897	0.8868
GNED	0.899	0.893	0.882	0.887	0.898	0.8918
NCEL	0.878	0.879	0.889	0.87	0.89	0.8812
E2ENEL	0.892	0.866	0.873	0.882	0.894	0.8814
BOOTLEG	0.891	0.891	0.874	0.865	0.873	0.8788
EDEGE	0.924	0.921	0.914	0.908	0.919	0.9172

To distinguish the proposed EDEGE method from baseline methods, the maximum value is bold in each column.