Context Attention Heterogeneous Network Embedding

Table 3

AUC scores on Cora.

% training edges	15%	25%	35%	45%	55%	65%	75%	85%	95%
DeepWalk	0.614	0.708	0.777	0.807	0.853	0.858	0.871	0.877	0.898
LINE	0.608	0.743	0.807	0.827	0.853	0.865	0.870	0.885	0.894
Node2vec	0.654	0.722	0.768	0.812	0.838	0.861	0.871	0.878	0.908
GraRep	0.589	0.732	0.786	0.826	0.852	0.874	0.897	0.898	0.914
Naive Combination	0.668	0.772	0.801	0.826	0.852	0.866	0.904	0.921	0.942
TADW	0.803	0.824	0.834	0.862	0.887	0.888	0.903	0.918	0.945
TENE	0.779	0.818	0.822	0.859	0.879	0.881	0.892	0.913	0.916
ASNE	0.718	0.742	0.809	0.832	0.849	0.870	0.902	0.921	0.933
CAHNE(w/o context)	0.654	0.747	0.803	0.843	0.877	0.885	0.901	0.909	0.915
CAHNE	0.793	0.805	0.828	0.863	0.892	0.898	0.908	0.925	0.954
CAHNE-a	0.805	0.830	0.837	0.872	0.892	0.907	0.915	0.926	0.963