Effectively Detecting Communities by Adjusting Initial Structure via Cores

Table 4

A comparison of the results of applying different similarity measures to ACC algorithm.

Network	Measures		CN	Jaccard	Salton
Risk Map	Q	0.634	0.600	0.600	0.631
	Accuracy	1.000	0.881	0.881	0.833
	NMI	1.000	0.857	0.857	0.871
	ARI	1.000	0.735	0.735	0.734
	Threshold	0	1	0.1	0.4
Dolphins	Q	0.520	0.520	0.512	0.510
	Accuracy	0.984	0.984	0.887	0.903
	NMI	0.949	0.949	0.881	0.839
	ARI	0.945	0.945	0.808	0.796
	Threshold	0	2	0.2	0.4
Karate	Q	0.371	0.371	0.371	0.371
	Accuracy	1.000	1.000	1.000	1.000
	NMI	1.000	1.000	1.000	1.000
	ARI	1.000	1.000	1.000	1.000
	Threshold	0	8	0.3	0.5
Polbooks	Q	0.495	0.416	0.441	0.486
	Accuracy	0.810	0.790	0.809	0.714
	NMI	0.535	0.530	0.519	0.506
	ARI	0.641	0.564	0.543	0.568
	Threshold	0	6	0.1	0.6
Football	Q	0.560	0.579	0.555	0.587
	Accuracy	0.852	0.809	0.852	0.817
	NMI	0.952	0.916	0.944	0.914
	ARI	0.853	0.805	0.850	0.734
	Threshold	0	2	0.3	0.3
Santa Fe	Q	0.718	0.724	0.709	0.711
	Accuracy	0.847	0.856	0.864	0.873
	NMI	0.886	0.877	0.855	0.858
	ARI	0.827	0.816	0.791	0.802
	Threshold	0	1	0.8	0.9