Financial Fraud Detection on IEEE CIS
97AUROCgraph representation learning and risk discrimination framework
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| graph representation learning and risk discrimination framework2026.05 | 97 | 93 | 92 | 91 | 91 | 96 | 2 | 7 | |
| LightGBMTrack=A, Model Type=Transaction-level2026.05 | 96.2 | — | — | — | 72.6 | 77.2 | — | — | |
| Tian et al.2026.05 | 95 | 90 | 89 | 88 | 88 | 93 | 8 | 16 | |
| Chen et al.2026.05 | 95 | 91 | 90 | 89 | 89 | 94 | 5 | 12 | |
| Liu et al.2026.05 | 94 | 89 | 88 | 87 | 87 | 92 | 9 | 14 | |
| Qian et al.2026.05 | 94 | 89 | 88 | 87 | 87 | 92 | 6 | 8 | |
| XGBoostTrack=A, Model Type=Transaction-level2026.05 | 93.2 | — | — | — | 61.6 | 65.1 | — | — | |
| Dou et al.2026.05 | 93 | 88 | 87 | 86 | 86 | 91 | 11 | 17 | |
| Tong et al.2026.05 | 93 | 88 | 87 | 86 | 86 | 92 | 8 | 10 | |
| Cheng et al.2026.05 | 92 | 87 | 86 | 85 | 85 | 90 | 13 | 18 | |
| Random ForestTrack=A, Model Type=Transaction-level2026.05 | 89.8 | — | — | — | 55 | 56.1 | — | — | |
| BiLSTMTrack=A, Model Type=Transaction-level, Stage=42026.05 | 80.7 | — | — | — | 44.3 | 38.2 | — | — | |
| SilIFalpha=0.5, number of seeds=5, trees=100, K=82026.05 | 72.35 | — | — | — | — | 13.15 | — | — | |
| Isolation Forestalpha=0, trees=100, number of seeds=52026.05 | 72.2 | — | — | — | — | 12.59 | — | — | |
| SilIFalpha=1.0, number of seeds=5, trees=100, K=82026.05 | 71.97 | — | — | — | — | 13.39 | — | — | |
| Global K-MeansK=8, number of seeds=52026.05 | 71 | — | — | — | — | 14.48 | — | — | |
| HBOSbins=20, number of seeds=52026.05 | 67.86 | — | — | — | — | 7.35 | — | — | |
| ECODparameter-free=true, number of seeds=52026.05 | 65.32 | — | — | — | — | 6.52 | — | — |