基于支持向量数据描述的网络可生存性综合评估  

Comprehensive evaluation for network survivability based on support vector data description

作　　者：梁霄[1] 孟相如[1] 陈铎龙[1] 庄绪春[1] 

机构地区：[1]空军工程大学信息与导航学院,西安710077

出　　处：《计算机应用研究》2013年第3期853-855,860,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61003252);全军军事学研究生课题(2011JY002-524)

摘　　要：针对网络可生存性综合评估方法中指标权重难以确定的问题,提出了基于支持向量数据描述(SVDD)的网络可生存性综合评估方法。该方法分析了SVDD的几何意义,采用二进制粒子群(BPSO)算法对建立的评估特征指标集进行特征选择,将所得的特征指标集视为整体来建立SVDD分类模型,并以测试样本点与模型的相对距离为依据评估系统的可生存性,避免了综合评估中指标权重确定的主观性。最后通过网络实例验证了评估模型的有效性。Aimed at the problem of indexes weighting ensuring in comprehensive evaluation, this paper proposed a comprehen- sive evaluation for network survivability based on support vector data description （SVDD）. It analyzed the meaning of SVDD firstly, and constructed the survivability feature index set from the angle of key attributes and characteristics. Then the index set was selected and optimized by using the binary particle swarm optimization （BPSO） algorithm. This method constructed the SVDD classify model by regarding the optimum feature subset as a integrity. And it introduced the relative distance in kernel space between diagnostic samples and distributed spheres as a basis to evaluate the system survivability. It avoided the subjec- tivity of indexes weighting ensuring in comprehensive evaluation by using this method. Finally, it provided a network example to verify the validity of the evaluation method.

关 键 词：网络可生存性 综合评估 支持向量数据描述 二进制粒子群算法 特征选择 相对距离 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]