Table 10. Comparison of recent approaches for intrusion detection on the NSL-KDD dataset (training time on whole dataset while testing time on single data sample).
| Study | Method | Performance measures (%) | Multi-class | No of features | Time required (s) | Model size | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Feature selection | Classifier | Accuracy | Precision | Recall | F1-score | Training time | Testing time | ||||
| Proposed | RF-RFE | ML ensemble | 99.53 | 99.79 | 99.78 | 99.29 | ✓ | 13 | ~18 | 0.003 | ~242 kb |
| CNN | 95.04 | 95.13 | 95.11 | 95.02 | ✓ | ~390 | 0.16 | ~1,024 kb | |||
| RNN | 89.3 | 88.10 | 89.12 | 88.19 | ✓ | ~360 | 0.085 | ~225 kb | |||
| LSTM | 91.21 | 91.1 | 91.2 | 91.23 | ✓ | ~800 | 0.084 | ~1,240 kb | |||
| Otair et al. (2022) | GWO+PSO | KNN + SVM | 98.97 | – | – | – | X | 20 | ~1,680 | 0.15 | – |
| Roy et al. (2022) | – | RF | 98.5 | – | – | – | ✓ | – | ~454 | ~0.0030 | – |
| Gu & Lu (2021) | k-Best | RF + XGB + DT | 99.9 | 99.8 | 99.9 | 99.9 | X | 20 | ~8.21 | 0.0055 | – |
| Pokharel, Pokhrel & Sigdel (2020) | CNN | AE | 85.51 | 97.62 | 68.90 | – | X | – | ~1,800 | 0.054 | – |