Risk Assessment of Cyberbullying Based on Improved Bayesian Networks

In the digital age, the issue of network security has grown increasingly prominent and poses a severe threat to the health and order of cyberspace. By employing methods such as questionnaire, literature review, and case analysis to determine the influencing factors, a well-structured system is established to assess the risk. In this paper, based on traditional Bayesian networks, by optimizing the structure of Bayesian networks, enhancing the selection and estimation of parameters, and increasing the efficiency of computational reasoning, a more comprehensive risk assessment system is established. The results indicate that the probability of detecting the issue of network security through Bayesian network forward reasoning is approximately 56%. Through this study, the dynamic assessment of network security problem has played an effective role in monitoring and prevention, and has a positive impact on maintaining social security.