Skip to main content

Table 7 Experimental results

From: CT-GCN+: a high-performance cryptocurrency transaction graph convolutional model for phishing node classification

Experimental model

F1

ACC

AUC

BaF

0.8077

0.8172

0.8178

BaF + TmF

0.8963

0.8982

0.8985

BaF + TsF

0.8575

0.8603

0.8606

BaF + TaF

0.8288

0.8355

0.8360

BaF + TmF + TsF

0.8919

0.8943

0.8946

BaF + TmF + TaF

0.8901

0.8916

0.8919

BaF + TsF + TaF

0.8487

0.8525

0.8528

BaF + TmF + TsF + TaF

0.8883

0.8903

0.8906

BaF + TmF + Node2vec1

0.9497

0.9316

0.9113

BaF + TmF + Node2vec2

0.9295

0.9055

0.8908

BaF + TmF + T-EDGE

0.9498

0.9317

0.9154

Time information section + SMOTE

0.9503

0.9686

0.9485

Time information section + Graph_SMOTE

0.8422

0.9547

0.9445

CT-GCN+ (our model)

0.9507

0.9722

0.9667