Table 4.

Performance of methods MTCOV, NC and CESNA on three datasets, according to two different measures used in the Eq. (16).

Method	F1-score			Jaccard similarity
	facebook	football	polblogs	facebook	football	polblogs
MTCOV	$\mathbf{0.5}\pm \mathbf{0.1}$	$\mathbf{0.86}\pm \mathbf{0.03}$	$0.8\pm 0.2$	$\mathbf{0.4}\pm \mathbf{0.1}$	$\mathbf{0.82}\pm \mathbf{0.04}$	$0.8\pm 0.2$
NC	$0.48\pm 0.08$	$0.82\pm 0.06$	$\mathbf{0.95}\pm \mathbf{0.09}$	$0.36\pm 0.08$	$0.75\pm 0.08$	$\mathbf{0.9}\pm \mathbf{0.1}$
CESNA	$0.46\pm 0.09$	$0.7\pm 0.0$	$0.6\pm 0.0$	$0.33\pm 0.08$	$0.6\pm 0.0$	$0.4\pm 0.0$

The results are averages and standard deviations over ten independent runs and the best outcomes are bolded.