Enhancing network security with memory-augmented visual attention networks and predator–prey optimization models

Enhancing network security with memory-augmented visual attention networks and predator–prey optimization models

연구 분야: Analysis

학회: Iran Journal of Computer Science

The Network Intrusion Detection System (NIDS) plays a crucial role in identifying security threats that can compromise data integrity within networks and critical business infrastructure. Current research faces challenges in achieving a high recall rate while maintaining an acceptable analysis time, particularly when dealing with large-scale network traffic data. To address these issues, this study introduces a novel combined model that integrates the Memory-Augmented Visual Attention Network (MAVAN) with the Golden Predator Search Algorithm (GoPSA), referred to as MAVAN-GoPSA, for effective intrusion detection and classification. MAVAN enhances feature extraction by incorporating visual attention and memory elements, allowing the model to focus on relevant features while preserving the contextual information necessary for accurate classification. Meanwhile, GoPSA improves the feature selection process through a search algorithm that balances exploration and exploitation to identify the most pertinent features for intrusion detection. The performance of the MAVAN-GoPSA model is evaluated against eight different techniques across three datasets: NSL-KDD, CICIoT-2023, and UNSW-NB 15, using metrics such as accuracy, precision, recall, and F1-score. The results demonstrate that this hybrid model outperforms existing state-of-the-art methods, achieving an impressive accuracy of up to 99.2% across several datasets. Thus, the authors present a practical and effective approach to NIDS that successfully addresses the limitations of previous methods, ultimately enhancing network security.