A Study on Graph Theory Application and Efficacy of Cybersecurity Situational Awareness in Industrial IoT System

Abstract

Abstract This paper proposes a network security situational awareness model based on graph theory, with the primary goal of improving industrial IoT system security. At the beginning of this paper, graph theory is explained in depth, the mutual transformation of directed and undirected graphs is proposed, the empowerment graph abstracted from practical problems is defined, matrix storage is used to realize graph storage, and an isomorphism function is proposed to realize isomorphism judgment of graphs. Based on the principles of graph theory, we develop a network security situational awareness model and suggest a network risk assessment system. This system utilizes risk indices for vulnerability, services, hosts, and networks and assesses the risk, threat, and posture of a specific asset. The efficacy of the cyber security situational awareness model is examined. The average precision rate, recall rate, and F1 value of this paper’s model reach 99.2%, 98.9%, and 97.05%, respectively. The performance of the recognition precision rate of different cyber-attack types is 1%~8% higher than that of the CN model. The leakage rate and false alarm rate of network attacks are 5.41% and 6.16%, respectively, and the overall accuracy rate reaches 95.48%. In terms of the running effect, the average absolute error and mean squared error of this paper’s model are 0.1302 and 0.2709, which are lower than other comparison models.