Ego-zones: non-symmetric dependencies reveal network groups with large and dense overlaps

Applied Network Science

Table 11 Properties of real-world networks

Network	n	m	Degree		CC	Strong-prominents		Weak-prominents		Modularity
			Max	Avg		Total	Percent	Total	Percent
artist	50515	819090	1469	32.438	0.137	8755	17.3	325	0.6	0.604
as-22july06	22963	48436	2390	4.218	0.230	2416	10.5	363	1.5	0.663
astro-ph	14845	119652	360	16.120	0.669	3291	22.1	3473	23.3	0.755
Brightkite	56739	212945	1134	15.012	0.173	14297	25.1	1961	3.4	0.660
com-amazon	334863	925872	549	5.529	0.396	75603	22.5	43161	12.8	0.926
com-dblp	317080	1049866	343	6.622	0.632	65090	20.5	29139	9.1	0.810
cond-mat	13861	44619	107	6.438	0.651	2241	16.1	3315	23.9	0.862
cond-2005	36458	171735	278	9.420	0.656	6372	17.4	8904	24.4	0.786
email-Enron	33696	180811	1383	21.463	0.509	5180	15.3	2591	7.6	0.584
facebook	4039	88234	1045	43.691	0.605	61	1.5	1242	30.7	0.835
ChCh-Miner	1510	48512	443	64.254	0.304	197	13.0	48	3.1	0.392
new_sites	27917	205964	678	14.776	0.294	7276	26.0	1588	5.6	0.611
power	4941	6594	19	2.669	0.080	1494	30.2	144	2.9	0.935
PP-Decagon	19065	715602	251	75.069	0.233	3500	18.3	147	0.7	0.445
PP-Pathways	21521	338625	213	31.812	0.124	3492	16.2	46	0.2	0.386
Yeast	2224	6609	64	6.339	0.125	563	25.3	38	1.7	0.587
LFR _{20 500 2000}	10000	102054	200	20.411	0.399	1499	15.0	225	2.3	0.659
LFR _{7 60 4000}	10000	32262	99	6.452	0.349	2280	22.8	472	4.7	0.578

The collaboration networks and networks with ground-truth communities differ from others (except facebook) in weakly-prominent nodes. For networks of these two types, there is a higher percentage of weakly-prominent nodes than for other networks. Biological networks, social networks (except facebook), and communication networks have a lower clustering coefficient