Feature selection with genetic algorithm and particle swarm optimization from intrusion detection data

With the rapid advancement of technology and the widespread adoption of online services, ensuring the security of individuals and organizations within the digital environment has become a paramount concern. In this regard, the analysis of network traffic for the detection of cyberattacks represents a critical area of research within the domain of information security. This study investigates the efficacy of feature selection using Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Artificial Bee Colony (ABC) algorithms applied to network traffic data. The analysis focuses on binary classification, categorizing data as either “attack” or “normal” without providing granular details on specific attack types. Feature selection was performed on the USB_IDS_1 and CSE-CIC-IDS2018 datasets, with the k – Nearest Neighbor (k-NN) algorithm employed during the classification phase. Model performance was assessed using the F1-score metric. The experimental findings indicate that GA and PSO achieved favorable outcomes in feature selection.