Table 5 The performance of LLMs based-data augmentation on the NF-CSE-CIC-IDS2018-v2 dataset.
| Â | Our | CTGAN | SMOTE | ADASYN | Oversample | Undersample | Original |
|---|---|---|---|---|---|---|---|
DDOS attack-HOIC | 0.998 | 0.998 | 0.705 | – | 0.625 | 0.543 | 0.998 |
DoS attacks-Hulk | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.996 | 0.995 |
DDoS attacks-LOIC-HTTP | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.979 | 1.0 |
Infilteration | 0.471 | 0.467 | 0.03 | – | 0.03 | 0.031 | 0.476 |
SSH-Bruteforce | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.877 | 1.0 |
Bot | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.991 | 1.0 |
DoS attacks-GoldenEye | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.963 | 0.994 |
FTP-BruteForce | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.919 | 1.0 |
DoS attacks-SlowHTTPTest | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.930 | 1.0 |
DoS attacks-Slowloris | 1.0 | 1.0 | 1.0 | – | 1.0 | 0.759 | 0.667 |
Brute Force -Web | 0.205 | 0.247 | 0.01 | – | 0.009 | 0.007 | 0.028 |
DDOS attack-LOIC-UDP | 0.993 | 0.987 | 0.982 | – | 0.992 | 0.536 | 0.0 |
Brute Force -XSS | 0.163 | 0.0 | 0.003 | – | 0.003 | 0.002 | 0.0 |
SQL Injection | 0.079 | 0.0 | 0.002 | – | 0.002 | 0.001 | 0.028 |
Benign | 0.997 | 0.996 | 0.725 | – | 0.708 | 0.802 | 0.991 |
F1-Macro | 0.794 | 0.780 | 0.697 | – | 0.691 | 0.622 | 0.678 |
Accuracy | 0.995 | 0.992 | 0.591 | – | 0.568 | 0.668 | 0.984 |
