Research Article
Light Weighted CNN Model to Detect DDoS Attack over Distributed Scenario
Table 7
Comparison of online detection performance of different models under different time windows.
| Model name | Time window/s | Accuracy | False interception rate | Malicious traffic detection rate |
| LSTM-LightGBM | 70 | 0.85394 | 0.28498 | 0.85244 | 110 | 0.89384 | 0.20312 | 0.87258 | 190 | 0.87698 | 0.21015 | 0.88593 |
| LSTM-RF | 70 | 0.90894 | 0.20591 | 0.96852 | 110 | 0.92438 | 0.17419 | 0.95478 | 190 | 0.89394 | 0.19875 | 0.92574 |
| CNN-RF | 70 | 0.9569 | 0.17058 | 0.87244 | 110 | 0.95347 | 0.11789 | 0.88574 | 190 | 0.98397 | 0.20591 | 0.98657 |
|
|