Convolutional Neural Network Based Algorithm for Intrusion Detection in Internet of Things Databases

The Internet of Things (IoT) is generally isolated from external networks using firewall technology[1]. However, with the continuous development of IoT technology, the interconnection is becoming closer and the frequency of intrusions is increasing[2]. This poses greater challenges for security measures. Therefore, it is crucial to efficiently detect intrusions in the IoT. This article proposes a convolutional neural network-based algorithm for intrusion detection in IoT databases[3]. IoT data has fixed formats, real-time characteristics, and short fields. Therefore, training with a convolutional neural network improves efficiency[4]. The algorithm monitors and analyzes network traffic in the IoT system, uses a convolutional neural network to classify and determine traffic data[5], and thus achieves attack detection in the IoT system. The algorithm designed in this article achieves an accuracy rate of 99.4%, which not only improves the accuracy of intrusion detection compared to SVM and KNN algorithms but also reduces false positive rates.