Levy flight salp swarm algorithm-based feature selection method for network intrusion detection systems

• Published in: AIP conference proceedings Vol. 2400; no. 1
• Main Authors: Saleh, Hadeel M., Hameed, Saif Saad, Abdulkareem, Ahmed B.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 31.10.2022
• Subjects: Algorithms; Classification; Datasets; Feature selection; Intrusion detection systems; Redundancy
• DOI: 10.1063/5.0112538
• ISSN: 0094-243X

One of the primary issues in this subject is the low accuracy of existing Network Intrusion Detection Systems (IDS); this issue is exacerbated by the high dimensionality of the feature selection process prior to the creation of IDS models. This challenge is typically handled by employing feature selection techniques in order to reduce dataset redundancy and improve classification performance. relevant subset of features, and reduce data dimensionality. Every Salp in the people was symbolized in binary form in this method, with 1 representing a selected feature and 0 representing a non-selected feature. The suggested feature selection approach was tested using the NSL-KDD dataset, which has 41 features. The final result of the, where 1 denotes a selected feature and 0 denotes a feature that is not selected. The suggested feature selection approach was tested using the NSL-KDD dataset, which has 41 features. The study revealed that the proposed strategy can increase the number of features picked and enhance classification accuracy.