List your company
ResearchFormerly WizWand

Graph Anomaly Detection on Amazon (AUC, F1-Macro)

98.89AUCSignGAD
62.843672.201881.5690.9182May 26, 2026
Updated 5mo ago

Evaluation results

MethodLinks
SignGADTraining Ratio=10%2026.0598.8993.51
SignGADTraining Ratio=5%2026.0598.6693.19
SignGADTraining Ratio=3%2026.0598.2392.54
TAQ-GADTraining Ratio=10%2026.0597.8292.59
TAQ-GADTraining Ratio=5%2026.0597.4392.14
SignGADTraining Ratio=1%2026.0597.4191.17
SignGADtraining ratio=1%, LLM backbone=GPT-4.12026.0597.4191.17
TAQ-GADTraining Ratio=3%2026.0596.8691.83
SpaceGNNTraining Ratio=10%2026.0596.4492.76
ConsisGADTraining Ratio=10%2026.0596.1792.11
SpaceGNNTraining Ratio=5%2026.0596.0892.41
ConsisGADTraining Ratio=5%2026.0595.8291.88
TAQ-GADTraining Ratio=1%2026.0595.7190.22
TAQ-GADtraining ratio=1%2026.0595.7190.22
SpaceGNNTraining Ratio=3%2026.0595.2991.57
ConsisGADTraining Ratio=3%2026.0595.1491.26
SpaceGNNTraining Ratio=1%2026.0594.0290.69
SpaceGNNtraining ratio=1%2026.0594.0290.69
ConsisGADTraining Ratio=1%2026.0593.9190.03
ConsisGADtraining ratio=1%2026.0593.9190.03
GDNtraining ratio=1%2026.0592.4389.75
MLPtraining ratio=1%2026.0592.3987.53
PC-GNNtraining ratio=1%2026.0591.1885.25
GraphSAGEtraining ratio=1%2026.0590.1284.25
CARE-GNNtraining ratio=1%2026.0589.6875.74
BWGNNtraining ratio=1%2026.0588.5690.48
GHRNtraining ratio=1%2026.0588.3586.35
GCNtraining ratio=1%2026.0587.3470.94
DiG-In-GNNtraining ratio=1%2026.0586.382.61
GATv2training ratio=1%2026.0585.3976.2
GINtraining ratio=1%2026.0584.3571.2
H2-Fdetectortraining ratio=1%2026.0583.8172.46
GAGAtraining ratio=1%2026.0582.6176.85
GATtraining ratio=1%2026.0580.7463.45
GraphConsistraining ratio=1%2026.0564.2355.35