Table 2 Comparison of prediction accuracy under the metric of root mean squared error for the top-L ranked links.

Network\Index	CN	WCN	rWCN	AA	WAA	rWAA	RA	WRA	rWRA
Celegans	0.535	0.529	0.531	0.528	0.524	0.527	0.531	0.53	0.536
Celegans	0.0587	0.0595	0.0583	0.0623	0.0629	0.0627	0.0616	0.0616	0.0607
Everglades	0.0821	0.0795	0.0978	0.089	0.107	0.104	0.101	0.116	0.11
Everglades	0.0686	0.0806	0.124	0.082	0.118	0.135	0.102	0.127	0.132
USAir1	0.00217	0.00359	0.00419	0.00223	0.00426	0.00433	0.00216	0.00419	0.00441
USAir1	0.00194	0.00485	0.00468	0.00197	0.00455	0.00477	0.0018	0.00445	0.00466
USAir2	0.73	0.73	0.731	0.729	0.729	0.73	0.726	0.725	0.724
USAir2	0.045	0.0446	0.0442	0.0455	0.045	0.0447	0.053	0.0529	0.0499
Advogato	0.313	0.311	0.309	0.311	0.308	0.306	0.313	0.311	0.307
Advogato	0.0408	0.0416	0.0424	0.0439	0.0457	0.0466	0.0506	0.0518	0.0528
Geom	0.435	0.419	0.41	0.39	0.377	0.37	0.369	0.358	0.347
Geom	0.0786	0.077	0.0767	0.11	0.105	0.0978	0.129	0.131	0.12

In each network, the first row is the results achieved by the linear-correlation method, while the second row shows the accuracy of our method. Each accuracy value is an average over 100 independent random divisions of the links into a training set and a validation set.

Quick links

Search