Identification of important nodes on large-scale Internet based on unsupervised learning

Abstract

In recent years, scholars have conducted in-depth researches on the robustness, structural vulnerability, and detection and identification of devices in cyberspace from different perspectives such as complex networks and cyberspace resource mapping. Aiming at the problem of identifying important nodes on a large-scale Internet, a CETCRank algorithm for identifying important Internet nodes based on unsupervised learning is proposed. When the algorithm analyzes the attributes of each cyberspace equipment, it not only considers the graph structure characteristics based on the network topology, but also integrates the threat metric of cyberspace equipment. Based on the hypothesis of the cyber attack model, the effective identification of important nodes in the Internet can be realized by integrating the node attributes into the constructed Markov chain model. Experiments show that the time and space complexity of the CETCRank algorithm is suitable for analyzing large-scale Internet, and the recognition performance of important nodes is better than the PageRank algorithm.