Research Article
A Residual Learning-Based Network Intrusion Detection System
Table 5
Comparison of results of recall with other classifiers.
| Model | Testing metric R | Analysis | Backdoors | DoS | Exploits | Fuzzers | Generic | Normal | Recon | Shellcode | Worms |
| RF | 0 | 0 | 0.012 | 0.839 | 0.741 | 0.630 | 0.999 | 0.754 | 0.111 | 0 | SVM | 0.062 | 0.033 | 0.183 | 0.669 | 0.730 | 0.763 | 0.997 | 0.645 | 0.206 | 0.068 | MLP | 0.007 | 0.039 | 0.076 | 0.825 | 0.679 | 0.891 | 0.999 | 0.661 | 0.508 | 0.227 | LSTM | 0 | 0 | 0.008 | 0.853 | 0.793 | 0.747 | 0.998 | 0.683 | 0 | 0.045 | LeNet-5 | 0 | 0 | 0 | 0.852 | 0.689 | 0.901 | 0.996 | 0.787 | 0.553 | 0 | AlexNet | 0 | 0.002 | 0.057 | 0.833 | 0.725 | 0.915 | 0.999 | 0.711 | 0.352 | 0.114 | RLC-CNN | 0 | 0.026 | 0.077 | 0.875 | 0.711 | 0.926 | 0.998 | 0.805 | 0.370 | 0 | RLF-CNN | 0.435 | 0.482 | 0.311 | 0.860 | 0.732 | 0.918 | 0.993 | 0.797 | 0.712 | 0.818 |
|
|