Research Article
Network Intrusion Detection Method Based on FCWGAN and BiLSTM
Table 8
Ablation experiment detection rate of various types.
| Algorithm | Type of samples | Normal | DoS | Probe | U2R | R2L |
| BiLSTM | 94.65 ± 0.21 | 88.24 ± 0.19 | 72.91 ± 0.23 | 46.81 ± 0.35 | 51.97 ± 0.30 | GAN-BiLSTM | 95.31 ± 0.25 | 92.18 ± 0.22 | 81.27 ± 0.30 | 60.33 ± 0.42 | 65.10 ± 0.37 | CWGAN-BiLSTM | 98.54 ± 0.19 | 94.60 ± 0.15 | 85.15 ± 0.23 | 70.20 ± 0.26 | 72.13 ± 0.25 | Model in this paper | 99.68 ± 0.14 | 96.01 ± 0.11 | 90.12 ± 0.15 | 76.35 ± 0.27 | 80.26 ± 0.19 |
|
|