Context Attention Heterogeneous Network Embedding

Table 4

AUC scores on Zhihu.

% training edges	15%	25%	35%	45%	55%	65%	75%	85%	95%
DeepWalk	0.469	0.472	0.497	0.507	0.533	0.537	0.556	0.574	0.587
LINE	0.521	0.569	0.618	0.624	0.655	0.636	0.646	0.676	0.698
Node2vec	0.488	0.482	0.507	0.505	0.552	0.546	0.558	0.582	0.590
GraRep	0.583	0.619	0.642	0.659	0.654	0.662	0.663	0.668	0.663
Naive Combination	0.524	0.553	0.579	0.618	0.653	0.672	0.689	0.705	0.703
TADW	0.558	0.576	0.593	0.625	0.655	0.697	0.696	0.723	0.729
TENE	0.551	0.549	0.607	0.622	0.660	0.666	0.668	0.692	0.711
ASNE	0.586	0.563	0.608	0.633	0.661	0.682	0.699	0.700	0.728
CAHNE(w/o context)	0.595	0.600	0.603	0.604	0.612	0.618	0.639	0.657	0.679
CAHNE	0.623	0.693	0.706	0.709	0.707	0.711	0.713	0.722	0.731
CAHNE-a	0.631	0.707	0.721	0.724	0.723	0.727	0.736	0.748	0.759