Research on Random Intrusion Depth Detection of Internet of Things Based on 3D Convolutional Neural Network

There are many problems in the industrial Internet of Things, such as low feature extraction rate, low detection efficiency and poor adaptability. To solve this problem, a random intrusion depth detection method based on three-dimensional convolution neural network is proposed. According to NIDS, an intrusion detection model of the Internet of Things is built, through which distributed network data packets are collected, and the principal component analysis algorithm is used to preprocess them to reduce data dimensions. Combined with deep learning theory and technology, select data features to form feature matrix. With this as the input, the random intrusion detection in the Internet of Things is completed by using 3D convolution neural network (3DCNN) combined with long and short memory (LSTM) method. The experimental results show that the F1 value of the detection method is above 0.9, indicating that the detection accuracy of the method is high.