Feature Selection Techniques Using Improved Bacterial Forage Optimization Algorithm for Network Intrusion Detection System
Main Article Content
Abstract
The increasing interconnections and accessibility of computing devices have become essential for improving our day-to-day operations. Distributed denial of services (DDoS) assaults on networking are detected by network intrusion detection technologies (NIDS), which show up as abrupt and large spikes in traffic across networks. These attacks seek to interfere with the availability of individual nodes or the system through resource depletion or signal to jam. In response to these difficulties, a novel method for detecting and categorizing DDoS attacks is used: bacterial foraging optimization with random forest Classifier optimization. The approach works in the databases CIDD and UGR16 to preprocess the incoming data using an autoencoder. This benchmark dataset works well for studies comparing various intrusion detection techniques. Then, the proposed BFO-RF optimization strategy partitions the data, emphasizing low-rate assaults. A random forest classifier is used to classify assaults once the feature selection procedure is completed. The effectiveness of the implemented BFO-RF optimization method is assessed, and a remarkable accuracy of 99.91% is obtained. By comparison, the accuracy of the well-known spider monkey optimized with hierarchical swarm optimization of particles (SMO-HPSO), Firefly Swarm Optimization, and Cuckoo Search Optimization techniques acquired an accuracy rate of 98.17. The result analysis shows unequivocally that the suggested BFO-RF optimization strategy is significantly more reliable than the current techniques. Because of its effectiveness, the proposed method has the ability to resolve practical optimization issues that arise across a variety of application areas; BFO algorithm has already caught the interest of researchers in various applications.
Article Details
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.
References
The increasing interconnections and accessibility of computing devices have become essential for improving our day-to-day operations. Distributed denial of services (DDoS) assaults on networking are detected by network intrusion detection technologies (NIDS), which show up as abrupt and large spikes in traffic across networks. These attacks seek to interfere with the availability of individual nodes or the system through resource depletion or signal to jam. In response to these difficulties, a novel method for detecting and categorizing DDoS attacks is used: bacterial foraging optimization with random forest Classifier optimization. The approach works in the databases CIDD and UGR16 to preprocess the incoming data using an autoencoder. This benchmark dataset works well for studies comparing various intrusion detection techniques. Then, the proposed BFO-RF optimization strategy partitions the data, emphasizing low-rate assaults. A random forest classifier is used to classify assaults once the feature selection procedure is completed. The effectiveness of the implemented BFO-RF optimization method is assessed, and a remarkable accuracy of 99.91% is obtained. By comparison, the accuracy of the well-known spider monkey optimized with hierarchical swarm optimization of particles (SMO-HPSO), Firefly Swarm Optimization, and Cuckoo Search Optimization techniques acquired an accuracy rate of 98.17. The result analysis shows unequivocally that the suggested BFO-RF optimization strategy is significantly more reliable than the current techniques. Because of its effectiveness, the proposed method has the ability to resolve practical optimization issues that arise across a variety of application areas; BFO algorithm has already caught the interest of researchers in various applications.