Metaheuristically Enhanced ANN-Based Intrusion Detection System with Explainable AI Integration

As smart devices continue to shape our lives and work, IoT networks have become an integral part of our daily lives. From smart homes to connected healthcare, these networks drive innovation but also come with a growing vulnerability, such as the risk of cyberattacks. Traditional intrusion detection systems often struggle to compensate for the complexity and sophistication of these threats, leaving critical security gaps. To address this challenge, we developed a robust ensemble approach to intrusion detection. First, we trained our dataset using a Convolutional Neural Network (CNN) and Artificial Neural Network (ANN). To further refine the performance of the ANN, we employed a metaheuristic optimization technique to ensure greater accuracy and reliability. Finally, we combine the strengths of both models using a stacking classifier to create an ensemble system capable of delivering exceptional results. The ensemble model achieved an impressive accuracy of 99.7%, outperforming the individual models, and offering a significant step forward in securing IoT networks. To make the system more transparent and trustworthy, we used Explainable AI (XAI) techniques, such as SHAP, allowing us to interpret the model’s decisions clearly. By blending innovation with usability, our approach not only advances intrusion detection but also inspires confidence in its ability to protect IoT networks that we rely on every day.