Deep Learning Based Network Intrusion Detection System Using Predator Optimization algorithm for Feature Selection

Network security is a critical concern in the digital era, necessitating effective intrusion detection systems (IDS) to protect against sophisticated cyber threats. This study presents a deep learning-based network intrusion detection system (NIDS) that is improved by using Predator Optimization Algorithm (POA) for optimal feature selection. Traditional IDS methods often face challenges with high-dimensional data, leading to increased computational costs and decreased detection performance. The integration of POA mitigates these issues by reducing the feature space and selecting the most relevant features for the learning process. Our system uses a deep learning architecture to detect intricate patterns and anomalies in network traffic data. The POA iteratively refines the feature subset, improving the model's performance. Experiments with the UNSW-NB15 datasets show that our method outperforms conventional techniques in accuracy, detection rate, and computational efficiency. The achieved results demonstrate that combining deep learning with POA for feature selection significantly enhances the IDS's ability to detect various intrusions accurately, prevent overfitting, and maintain low false positive rates. This approach offers a promising solution for developing advanced NIDS, enhancing security for critical network infrastructures.