摘要
嵌入式单片机系统软件安全漏洞检测效果的优劣不仅影响计算机的基本功能发挥,还会损害用户的切身利益。针对当前方法忽略了无关特征和冗余特征对嵌入式单片机系统软件安全漏洞检测的影响,导致检测结果不佳,提出一种基于朴素贝叶斯的软件安全漏洞自动检测方法,利用蚁群算法的搜索性能,将系统软件安全漏洞特征提取问题转化为路径寻优问题,计算蚂蚁的转移概率和适应度函数值,并对系统各个路径上的信息素浓度进行实时更新。采用额外附加激励的方式,强化蚂蚁对最优路径的选择影响。设置蚁群算法的终止条件,将搜索获得的软件安全漏洞特征输出。采用特征向量来描述输出的系统软件安全漏洞特征样本,根据嵌入式单片机系统软件安全漏洞特征的先验概率计算出其后验概率,能够使得后验概率获得最大值的类即为该安全漏洞特征对象所属的类。仿真测试结果表明,所提方法能够实现系统软件安全漏洞的分类检测,具有较高的检测率和较低的误报率,同时漏洞覆盖率评价指标最高可以达到97.7%,远远高于对比方法。
This paper presents a method to automatically detect software security vulnerability based on Naive Bayes. This method used search performance of ant colony algorithm to transform feature extraction of system software security vulnerability into the path optimization. Then, our method calculated the transition probability of ant and fit- ness function value and updated the pheromone concentration on each path of system in real time. Moreover, the method used additional excitation to strengthen the influence of ant on the selection of optimal path. In addition, our research outputted the software security vulnerability feature obtained by the search through the termination condition of ant colony algorithm. Finally, the research used the feature vector to describe the outputted sample of system soft- ware security vulnerability feature. According to the prior probability of security vulnerability feature of software in embedded single - chip microcomputer system, we calculated the posterior probability. Thus, posterior probability could obtain the class of the maximum values which the security vulnerability feature object belonged to. Simulation results show that the proposed method can realize the classified detection of system software security vulnerability, which has high detection rate and low false alarm rate. Meanwhile, the evaluation index of vulnerability coverage rate can reach 97.7% , which is higher than comparison method.
作者
袁钦
李利花
YUAN Qin;LI Li - hua(Gongqing College,Nanchang University,Gongqingcheng Jiangxi 332020,Chin)
出处
《计算机仿真》
北大核心
2018年第8期405-409,共5页
Computer Simulation
关键词
嵌入式
单片机系统
软件
安全漏洞
自动检测
Embedded
Single - chip microcomputer system
Software
Security vulnerability
Automatic detection
