基于云参数贝叶斯网络的威胁评估方法

An Threat Assessment Method Based on Cloud Parameters Bayesian Network

文中以威胁评估为背景,针对威胁评估中样本数据不充足,专家构建贝叶斯网络参数工作量大的问题,提出了基于云参数贝叶斯网络的威胁评估方法。把云的表达能力与贝叶斯网络的推理能力相结合,一是运用云的表达能力构建贝叶斯网络参数,二是运用贝叶斯网络的推理能力计算后验概率。首先,以状态组合权值为媒介运用专家知识构建隶属云模型,并利用状态组合权值的不确定度将隶属云模型转换为条件概率表,从而达到以较少的专家工作完成评估模型构建的目的;其次,运用专家构建的威胁评估贝叶斯网络和生成的条件概率表进行威胁评估推理,得到最终的评估结果。实验结果表明,该方法生成的条件概率表的统计数据与专家知识相符,并能有效地应用于威胁评估之中。

For the disadvantages of lacking sample data of threat assessment and large workload of experts building Bayesian network, a threat assessment method based on cloud parameters Bayesian network is proposed. The method combines cloud model expression ability with Bayesian network inference ability. On the one hand,the cloud expression ability is used to build a Bayesian network parameters,on the other hand, the Bayesian network inference ability is applied to calculate the posterior probability. First, it uses expert knowledge to generate membership cloud parameters with the media of state combination weight and converts membership cloud to conditional proba- bility tables by the uncertainty of state combination weights, so as to achieve the purpose to build the assessment model in less workload of experts. Then use of Bayesian network of threat assessment built by experts and conditional probability table generated for threat assessment reasoning, the final evaluation results are obtained. The experiment shows that this method is generated in line with the experts expected, and can be effectively applied to threat assessment.