Variational Approach for Learning Community Structures

Table 3

Experimental results on PolBlogs dataset.

	NMI ()	VI ()	ACC ()	Q ()	CON ()	TPR ()
Spectral Clustering	0.0014	1.1152	0.4828	-0.0578	0.5585	0.7221
Louvain	0.6446	1.0839	0.9149	0.2987	0.8130	0.1922
DeepWalk	0.7367	1.0839	0.9543	0.0980	0.3873	0.6870
node2vec	0.7545	0.8613	0.9586	0.1011	0.3827	0.6863
Stochastic Blockmodel	0.0002	1.2957	0.4905	-0.0235	0.5329	0.5657
Stochastic Blockmodel (D.C)	0.7145	0.8890	0.9496	0.4256	0.0730	0.8101
VGAE∗ + -means	0.7361	0.8750	0.9552	0.4238	0.0752	0.8089
VGAECD∗	0.7583	0.8583	0.9601	0.4112	0.0880	0.7913