Table 2.

Sample results of comparison between Algorithm 1 and local link prediction methods on real networks (full results in SI). We report top-precision obtained using 1000 test runs in the case of small networks and 100 test runs for larger networks. In each test run 10% of the edges are removed and used as test set. For every network, the results having the best significant rank with $p=0.05$ are shown in bold. The columns $n$ and $m$ show the number of nodes and edges of the networks.

Network	$n$	$m$	ADA	CNE	CH	HPI	JID	PAT	RAL	ALG1
Terrorist	62	152	0.2923	0.2443	0.2762	0.1568	0.0549	0.1059	0.2907	0.2936
SFBD Food Web	128	2,106	0.0724	0.0700	0.0860	0.0566	0.0113	0.1654	0.0725	0.1215
E.Coli	418	519	0.0243	0.0088	0.0009	0.0012	0.0003	0.0118	0.0278	0.0334
Web Edu	3,031	6,474	0.4267	0.2273	0.1928	0.2675	0.0103	0.0121	0.4247	0.5032
Advogato	5,155	39,285	0.1329	0.1214	0.1548	0.0122	0.0023	0.0502	0.1390	0.1539
PGP	10,680	24,316	0.3110	0.2591	0.3136	0.0824	0.0107	0.0212	0.2831	0.3434
EAT	23,219	304,938	0.0665	0.0652	0.0911	0.0016	0.0003	0.0126	0.0413	0.0651
AS CAIDA 2007 11 05	26,475	53,381	0.0480	0.0444	0.0488	0.0004	0.0000	0.0327	0.0350	0.0534
Func-Func	46,027	106,510	0.0546	0.0395	0.0007	0.0060	0.0000	0.0497	0.0437	0.0799
Livemocha	104,103	2,193,083	0.0119	0.0116	0.0262	0.0002	0.0000	0.0152	0.0113	0.0307
Amazon	334,863	925,872	0.1677	0.1272	0.1636	0.0888	0.0345	0.0004	0.1584	0.1850
Twitter Follows	404,719	713,319	0.0017	0.0008	0.0017	0.0003	0.0000	0.0013	0.0021	0.0025