Abstract
With the rapid development of Internet of Things (IoT) technology, the number of network attack methods it faces is also increasing, and the malicious network traffic generated is growing exponentially. To identify attack traffic for the protection of IoT device security, attack detection has attracted widespread attention from researchers. However, current attack detection methods struggle to identify complex and variable attack methods, resulting in a high false positive rate. Additionally, feature redundancy and class imbalance in IoT traffic datasets also constrain detection performance. To address these issues, this paper proposes an attack detection method based on deep learning for IoT. Firstly, a genetic algorithm is used for feature selection; secondly, a cost-sensitive function is employed to address the scarcity of attack traffic in IoT; and finally, a combination of Convolutional Neural Networks and Long Short Term Memory Network is utilized to extract spatiotemporal information from the network. The results demonstrate that this method exhibits superior performance on two IoT benchmark datasets, effectively enhancing the performance of IoT attack detection.