Research on Industrial Control Protocol Clustering Algorithm Based on Convolutional Self-Encoder and Improved K-means

Le résumé fourni par la source

The increase in the number of private protocols used in industrial control systems brings challenges to network security maintenance work, among which, the classification and analysis of unknown protocols is a difficult point to overcome. In order to solve the classification problem of unknown protocols, an industrial control protocol clustering algorithm based on convolutional self-encoder and improved K-means is proposed. Firstly, a protocol feature extraction model based on the convolutional self-encoder model is designed, and the local relationship and spatial feature extraction of industrial control messages are realized by introducing the convolutional layer, pooling layer and activation function, and then an industrial control protocol clustering algorithm based on the improved K-means is used to complete the classification of the industrial control protocols, and finally, the method is verified by testing to achieve better results in the clustering of industrial control protocols.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.