针对现有网络安全态势评估方法具有信息来源单一、评估范围有限、模型训练与参数获取不易和时空开销较大等问题,提出了一种基于聚类分析评估网络安全态势的方法.首先构建主机上若干二级广义立方体,快速高效聚类融合主机多源异构非确定性信息源,生成主机的安全态势,并利用Dirichlet先验分布平滑后验概率,优化聚类分析结果;然后构建网络上的广义立方体,聚类融合网络上各主机的安全态势,逐步量化评估并生成网络当前安全态势.通过真实网络环境的实验,验证了所提出的方法在网络安全态势评估中的可行性和有效性.
A new network security assessment method was proposed based on cluster analysis. First, a number of second-level general cubes were constructed, into which multi-source and heterogeneous information were gathered, using the principle of Dirichlet's prior distribution optimizing the result of posterior probability. Then, a network general cube was built to fuse the security of each host, assessing the security of network quantitatively step by step. The feasibility and effectiveness of the proposed method in the network security assessment were verified by the experiments of real network environment.