X-CYBERCNN: Explainable network architecture for transparent cyber-attack prediction in critical manet infrastructures

The Internet of Things (IoT) connects billions of devices, improving daily life while increasing cyber-security risks. Recent years have seen a rapid rise in the number and complexity of cyber-attacks targeting IoT infrastructure. To address this challenge, this research proposes a cyber-attack prediction framework for IoT environments by integrating MANET attack patterns with the X-cyberCNN model for transparent attack detection in critical MANETs.The explainable architecture classifies network traffic into multiple attack categories. X-cyberCNN enhances CNN performance using a Gated Recurrent Unit (GRU) to select and rank important features, improving attack detection accuracy. The framework was validated using real-world IoT traffic datasets. Experimental results demonstrated efficient attack classification and identification of key attributes affecting model performance.The method analyzes network traffic features associated with security threats and compromised nodes. Feature vectors were generated from data collected across various IoT devices. Performance evaluation showed that the enhanced X-cyberCNN model achieved 99.9% prediction accuracy, while multi-stage cyber-attacks reached a 75% occurrence rate, requiring highly accurate real-time threat management.The proposed system offers high computational efficiency and scalability, making it suitable for large-scale IoT networks by enabling fast and accurate attack prediction, early vulnerability detection, and effective risk mitigation