From: The use of generative adversarial networks to alleviate class imbalance in tabular data: a survey
Paper - Novel vs Baseline - Sampling Method | AUC | Balanced Accuracy | F1 Score | Precision | Recall |
---|---|---|---|---|---|
A Deep Learning Based Method for Handling Imbalanced Problem in Network Traffic Classification | Â | Â | Â | Â | Â |
 Novel |  |  |  |  |  |
  ACGAN | 88.42% | 99.80% | 86.67% |  |  |
  BalanceCascade | 95.10% | 99.90% | 95.10% |  |  |
 Baseline |  |  |  |  |  |
  no treatment |  | 99.49% | 82.92% |  |  |
  SMOTE | 94.12% | 99.20% | 94.94% |  |  |
FLOWGAN:Unbalanced network encrypted traffic identification method based on GAN | Â | Â | Â | Â | Â |
 Novel |  |  |  |  |  |
  ACGAN |  | 99.10% | 99.10% | 97.99% | 89.95% |
 Baseline |  |  |  |  |  |
  no treatment |  | 89.95% | 89.68% | 99.11% | 97.94% |
  Oversampling |  | 97.94% | 97.96% | 90.00% | 99.10% |
GAN-based imbalanced data intrusion detection system | Â | Â | Â | Â | Â |
 Novel |  |  |  |  |  |
  GAN RF |  | 99.83% | 95.04% | 98.68% | 92.76% |
 Baseline |  |  |  |  |  |
  RF |  | 99.19% | 87.79% | 98.20% | 83.79% |
  SMOTE |  | 99.51% | 88.16% | 88.97% | 87.51% |
PacketCGAN: Exploratory Study of Class Imbalance for Encrypted Traffic Classification Using CGAN | Â | Â | Â | Â | Â |
 Novel |  |  |  |  |  |
  ACGAN |  | 99.51% | 99.47% | 99.36% | 99.58% |
 Baseline |  |  |  |  |  |
  GAN |  | 97.66% | 97.66% | 97.66% | 97.67% |
  No treatment |  | 97.97% | 97.66% | 97.59% | 97.75% |
  Oversampling |  | 98.89% | 98.91% | 98.92% | 98.89% |
  SMOTE |  | 97.69% | 97.10% | 97.51% | 97.89% |