Variational Approach for Learning Community Structures

Table 5

Experimental results on PubMed dataset.

	NMI ()	VI ()	ACC ()	Q ()	CON ()	TPR ()
Spectral Clustering	0.0382	1.4341	0.3261	0.0414	0.5645	0.4935
Louvain	0.1983	3.6667	0.0954	0.7726	0.1388	0.1592
DeepWalk	0.2946	1.7865	0.3101	0.5766	0.0499	0.2461
node2vec	0.1197	1.9849	0.2228	0.3501	0.3170	0.2269
Stochastic Blockmodel	0.0004	1.9340	0.3080	-0.1620	0.1038	0.1965
Stochastic Blockmodel (D.C.)	0.1325	2.0035	0.3118	0.5622	0.8121	0.2441
VGAE∗ + -means	0.2041	1.8096	0.3724	0.5273	0.1320	0.2898
VGAE + -means	0.1981	1.8114	0.2751	0.5297	0.1283	0.2900
VGAECD∗	0.1642	1.8320	0.1956	0.4966	0.1252	0.2692
VGAECD	0.3252	1.7056	0.4216	0.6878	0.1636	0.4827