基于贝叶斯网络的等级保护测评辅助方法

An Auxiliary Method of Classified Protection Evaluation Based on Bayes Network

当前在等级保护测评活动中,测评指标众多,为减少人工审核所耗费的时间和资源,引入了贝叶斯网络作为一种辅助方法。首先,该方法以现有机构的大量测评数据为基础构建贝叶斯网络模型,模型以测评指标为父节点,测评指标中的各检查点为子节点,通过工具Netica计算得出各测评指标和检查点的概率;然后,通过专家验证调整部分预设关系得出各测评指标和检查点基准概率;最后,以此准概率与测评中新得到的数据进行对比,可为等级保护测评活动提供辅助检验、预测的功能。

There are many evaluation indicators in the evaluation of classified protection of cybersecurity.To reduce the time and resources consumed by manual review,Bayesian network is introduced as an auxiliary method.Firstly,based on a large number of evaluation data of existing institutions,a Bayesian network model is built.This model takes evaluation indicators as parent nodes and each checkpoint in evaluation index as child node,and the probability of each evaluation indicator and check point is calculated by Netica.Then,the benchmark probabilities of each evaluation index and checkpoint are obtained by verifying and adjusting some presct relationships with experts.Finally,the quasi-probability is compared with the newly obtained data,which can provide the functions of auxiliary inspection and prediction for grade protection evaluation activities.