Table 3 AUC comparison results on ten datasets (ML, EN, SWN, CL, CM, IC, C2O, Mal, GPC and Drug).

From: A potential energy and mutual information based link prediction approach for bipartite networks

Dataset

ML

EN

SWN

CL

CM

IC

C2O

Mal

GPC

Drug

PA

.881

.788

.648

.773

.764

.823

.901

.591

.720

.880

CN

.871

.851

.730

.811

.801

.910

.990

.901

.812

.920

JC

.791

.880

.663

.821

.798

.850

.950

.901

.821

.910

CAR

.912

.867

.726

.940

.906

.916

.990

.910

.811

.901

CJC

.882

.867

.762

.960

.940

.925

.990

.910

.831

.910

CAA

.910

.851

.760

.968

.950

.942

1.00

.920

.831

.910

CRA

.921

.890

.772

.945

.955

.931

1.00

.910

.821

.930

BPR

.911

.891

.742

.943

.959

.920

.990

.901

.840

.920

CS

.831

.836

.761

.775

.882

.835

.960

.821

.801

.871

PLP

.930

.889

.936

.905

.960

.945

.965

.906

.849

.938

NMF

.891

.761

.692

.854

.846

.850

.990

.861

.702

.890

PMIL

.945

.901

.945

.940

.982

.938

.971

.921

.867

.945

  1. Each dataset is divided into training set (90%) and test set (10%) and results are computed by averaging over 1000 runs.