Network Security Risk Evaluation and Inducement Analysis Based on Bayesian Method
哲 阿部, Shuaichen Ye · 2024
With the extension of the cyberspace, network safety issue emerges gradually, especially in scientific and industrial fields. To evaluate the network security status rapidly and precisely, a Bayesian network (BN) with 4 layers and 28 nodes is proposed in this paper. The BN diverges with the network system risk as the child node, and 5 qualitative risk stages of the network are defined. By taking an enterprise-level network system as an example, prior and posterior probabilities of each nodes are determined via long-term observation and expert experience, respectively. A transformation between the qualitative risk evaluation from BN and the quantitative risk value calculation is given. By analyzing influence factors of the system risk status, it is found that disposing rapidity, CPU resource anomaly and malware implantation attack are top three factors causing potential network safety hazard.