|
| Reference | Methodology | Dataset | Result |
|
| [42] | GAs, PSO, and differential evolution (DE) | KDD CUP 99 | DE outperforms GAs and PSO |
| [43] | Particle swarm optimization classifiers with genetic-particle swarm optimization | KDD CUP 99 | PSO is better than the fuzzy clustering technique |
| [44] | EFS algorithm and TLBO methodologies | NSL-KDD | Accuracy of 99.95% |
| [45] | PSO, GA, and neural networks (RBF) | KDD CUP 99 | The study result demonstrated that the precision, recall, and accuracy for RBF were 84.11, 84.21, and 84.43, respectively, whereas the precision, recall, and accuracy for GA was 85.68, 85.32, and 87.47 and for PSO was 86.13, 85.78, and 88.13, respectively |
| [46] | Systematic review on bioinspired swarm used to improve ML-based IDS | — | The study focuses on reviewing the implementation of SI algorithms to optimize feature selection and weight/parameter optimization to reduce computational and space complexities in ML systems and boost the detection accuracy |
| [47] | Survey of swarm and evolutionary algorithms in IDS systems | Different datasets used in the literature | The survey demonstrated the application of SWEVO algorithms in IDS. Further, a demonstration of the performance of SWEVO algorithms over traditional ML and DL IDS systems is presented |
| [48] | Review of different dataset benchmarks used in IDS evaluation. Identifying the weaknesses of current datasets and highlighting new datasets that reflect current network complexities | KDD
UNM
ADFA-LD | The study demonstrated the features of KDD and UNM datasets used to validate IDS. They showed that KDD and UNM are lacking some important features in modern complex networks. Further, they showed that ADFA-LD is more comprehensive in representing modern network complexities |
|
