Hybrid Deep Learning framework-based intrusion detection system for the Internet of Things

The recent surge in Internet of Things (IoT) devices has brought about unique security challenges. These devices have limited resources and are vulnerable to attacks. As Internet use becomes more widespread, network security has become a top priority, and network intrusion detection has emerged as a crucial technology for protecting systems. However, the original data sources for intrusion detection often have high dimensionality and large volumes, which can impact efficiency and accuracy. To address this issue, we propose a hybrid convolutional recurrent neural network (HCRNN-IDS) based intrusion detection system to detect malicious packets in real time. By utilizing a newly developed benchmark Netflow-based dataset like NF-QU-NIDS, our system can efficiently detect twenty different network attacks. We have conducted empirical evaluations to validate the accuracy of our model, and the results demonstrate its effectiveness in real-world scenarios. Furthermore, we have compared our proposed system with state-of-the-art intrusion detection methods, proving that our system achieves a superior accuracy rate of 98.44%.