Hybrid Intrusion Detection Model for Classification of Network Attacks with GA-PSO Optimization Technique
Akhilesh Kumar Shrivas, Prakash Pathak · Cybernetics & Systems · 2025
Ensuring information security and protecting sensitive data from unauthorized access is a critical concern for all organizations. To address this, we employ an Intrusion Detection System (IDS) designed to identify and prevent unauthorized activities before they cause harm. IDS functions as a classifier, distinguishing normal data from potential threats. The primary objective of this research is to evaluate the effectiveness of our proposed hybrid model in terms of computational efficiency and robustness. This research work explores various Machine Learning (ML) based classification methods and develops a new ensemble model. This model, referred to as the Proposed Voting Ensemble Classifier (PVEC), integrates Extra Tree Classifier (ETC), Histogram Gradient Boosting Classifier (HGBC), and CatBoost Classifier (CTBC) using a voting-based ensemble technique. Additionally, we applied different metaheuristic optimization techniques like Particle Swarm Optimization (PSO), Genetic Algorithms (GA), and a hybrid GA-PSO approach to enhance feature selection and improve the computational efficiency of IDS. The proposed hybrid model (PVEC-GA-PSO) combines PVEC with GA-PSO to optimize the features of IDS dataset and computationally improve the performance of model. Our results demonstrate that proposed hybrid model achieves superior performance for classifying network attacks, attaining the highest accuracy, precision, and F1 score for both binary and multiclass classification tasks.