Table 2 The average predicting AUC obtained by 100 independent runs on 19 real undirected networks.

From: Link prediction in complex networks via matrix perturbation and decomposition

Network

CN

AA

RA

SPM

LR

MPD

Karate

0.692

0.728

0.736

0.784

0.561(0.23)

0.789(0.23)

Football

0.679

0.682

0.677

0.800

0.610(0.18)

0.828(0.12)

Dolphin

0.786

0.789

0.786

0.773

0.604(0.26)

0.840(0.30)

Everglades

0.681

0.692

0.701

0.939

0.928(0.12)

0.946(0.16)

WorldTrade

0.868

0.889

0.907

0.934

0.861(0.12)

0.935(0.33)

Macaca

0.945

0.945

0.948

0.985

0.949(0.17)

0.989(0.18)

FWM

0.710

0.712

0.716

0.926

0.873(0.13)

0.932(0.19)

BUP

0.886

0.895

0.898

0.889

0.658(0.17)

0.897(0.36)

WorldAdj

0.679

0.676

0.675

0.718

0.539(0.14)

0.758(0.06)

FWF

0.610

0.608

0.612

0.950

0.853(0.14)

0.951(0.28)

Jazz

0.956

0.964

0.973

0.976

0.868(0.13)

0.979(0.27)

Contact

0.959

0.962

0.961

0.960

0.863(0.10)

0.980(0.02)

C. elegans

0.847

0.862

0.869

0.894

0.562(0.10)

0.904(0.34)

USAir

0.953

0.965

0.972

0.942

0.841(0.08)

0.950(0.18)

INF

0.942

0.945

0.947

0.943

0.754(0.13)

0.946(0.31)

Metabolic

0.922

0.955

0.960

0.931

0.595(0.10)

0.937(0.38)

Email

0.856

0.858

0.857

0.899

0.562(0.16)

0.910(0.25)

PB

0.923

0.929

0.929

0.926

0.656(0.07)

0.965(0.05)

Yeast

0.915

0.917

0.917

0.969

0.878(0.14)

0.987(0.14)

  1. The training set contains 90% of total connections. The values in the brackets are the values of optimal parameters of the methods in the simulations. The highest precisions are shown in boldface.