基于模糊聚类分析的网络安全综合评价

模糊聚类分析主要用于研究样本的分类问题。利用层次分析方法给出了样本指标的权重，进一步引入了加权相似系数和加权距离系数的计算公式。在该系数的基础上，提出了一种既能考虑到样本之间的值贴近程度，又能考虑到样本之间的形贴近程度的改进统计量——相似度，用相似度矩阵替代传统的相似矩阵，得到了模糊聚类分析方法的改进。将其应用到网络安全综合评价实际样本数据的聚类分析中，结果表明改进的模糊聚类分析方法是有效的。

Comprehensive assessment of network security base on fuzzy clustering analysis

Fuzzy clustering is mainly applied in the sample classification. The weights of sample index are given by analytical hierarchy process, and expressions of weighted similar coefficient and weighted distance coefficient are induced. Based on the coefficients, similar degree has been presented which could describe not only the degree of value similarity, but also of shape similarity. Replaced conventional similar matrix by similar degree matrix, fuzzy clustering analysis method is improved. The use of this method in practical sample data of comprehensive assessing index system of network security is presented. The results show that this improved fuzzy clustering analysis method is effective.