Variational Approach for Learning Community Structures

Table 4

Experimental results on Cora dataset.

	NMI ()	VI ()	ACC ()	Q ()	CON ()	TPR ()
Spectral Clustering	0.0651	2.0005	0.1252	0.0189	0.1909	0.6196
Louvain	0.4336	4.0978	0.0081	0.8142	0.0326	0.2821
DeepWalk	0.3796	2.7300	0.1626	0.6595	0.0396	0.4949
node2vec	0.3533	2.9947	0.1359	0.6813	0.1078	0.4902
Stochastic Blockmodel	0.0917	3.5108	0.1639	0.4068	0.4280	0.3376
Stochastic Blockmodel (D.C.)	0.1679	3.4547	0.1176	0.6809	0.1736	0.5112
VGAE∗ + -means	0.2384	3.3151	0.1033	0.6911	0.1615	0.4906
VGAE + -means	0.3173	3.1277	0.1589	0.6981	0.1517	0.5031
VGAECD∗	0.2822	3.1606	0.1532	0.6674	0.1808	0.5076
VGAECD	0.5072	2.7787	0.1101	0.7029	0.1371	0.4987