Fraud Detection on Credit Card Fraud Detection (test)
91.83RecallRandom Forest
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Random Forest# Samples=50, Oversampler=VAE-GAN+CPAC2025.07 | 91.83 | 96.58 | 96.85 | 94.08 | — | |
| XGBoost# Samples=50, Oversampler=VAE-GAN+CPAC2025.07 | 91.83 | 94.55 | 98.49 | 93.15 | — | |
| Random Forest# Samples=75, Oversampler=VAE-GAN+CPAC2025.07 | 91.83 | 96.58 | 96.85 | 94.08 | — | |
| XGBoost# Samples=100, Oversampler=VAE-GAN+CPAC2025.07 | 91.83 | 94.55 | 98.89 | 93.15 | — | |
| XGBoost# Samples=75, Oversampler=VAE-GAN+CPAC2025.07 | 90.81 | 94.43 | 98.31 | 92.54 | — | |
| Random Forest# Samples=100, Oversampler=VAE-GAN+CPAC2025.07 | 90.81 | 96.5 | 96.81 | 93.47 | — | |
| KNN# Samples=50, Oversampler=VAE-GAN+CPAC2025.07 | 89.79 | 95.33 | 94.88 | 92.38 | — | |
| KNN# Samples=75, Oversampler=VAE-GAN+CPAC2025.07 | 89.79 | 95.33 | 94.88 | 92.38 | — | |
| KNN# Samples=100, Oversampler=VAE-GAN+CPAC2025.07 | 89.79 | 95.33 | 94.88 | 92.38 | — | |
| Logistic Regression# Samples=50, Oversampler=VAE-GAN+CPAC2025.07 | 89.77 | 86.25 | 97.05 | 87.48 | — | |
| Logistic Regression# Samples=100, Oversampler=VAE-GAN+CPAC2025.07 | 89.77 | 86.09 | 97.38 | 87.84 | — | |
| Logistic Regression# Samples=75, Oversampler=VAE-GAN+CPAC2025.07 | 88.75 | 85.83 | 97.59 | 87.23 | — | |
| Majority VoteBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 84.6 | — | — | 89.6 | 15.4 | |
| Best IndividualBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 84.4 | — | — | 88.5 | 15.6 | |
| BrierBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 83.9 | — | — | 89.6 | 16.1 | |
| ExternalityBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 83.9 | — | — | 89.6 | 16.1 | |
| Conf-WeightedBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 83.9 | — | — | 89.6 | 16.1 | |
| VCG (Ours)Base Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 83.8 | — | — | 88.1 | 16.2 | |
| Stacking-MLPBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=5, Stacking=3-fold cross-validation2026.05 | 83 | — | — | 89.4 | 17 | |
| XGBoostNumber of Top Features=182026.02 | 82.2 | 80.6 | 97.7 | 81.4 | — | |
| Stacking-LRBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=5, Stacking=3-fold cross-validation2026.05 | 82.1 | — | — | 89.1 | 17.9 | |
| Log-OddsBase Learners=LightGBM/CatBoost/XGBoost/RF/MLP, Threshold=τ=0.3, Feature Partition Size=60%, Seeds=10, Agents=52026.05 | 81.5 | — | — | 88.8 | 18.5 | |
| Decision TreeNumber of Top Features=182026.02 | 81 | 69.9 | 92.4 | 74.6 | — | |
| Logistic RegressionNumber of Top Features=182026.02 | 80.6 | 76.5 | 97.9 | 78 | — | |
| EBMNumber of Top Features=182026.02 | 76.3 | 91.7 | 98.3 | 83.1 | — | |
| Random ForestNumber of Top Features=182026.02 | 76.1 | 94.2 | 97.6 | 83.7 | — |