基于Fisher算法的本体推理安卓恶意软件安全检测  

Ontology reasoning of Android malware security detection based on Fisher classification algorithm

作　　者：曹怀宝 张红梅[1] CA;ZHANG Hongmei(Guangxi Colleges and Universities Key Laboratory of Cloud Computing and Complex Systems,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学广西高校云计算与复杂系统重点实验室,广西桂林541004

出　　处：《桂林电子科技大学学报》2018年第4期290-295,共6页Journal of Guilin University of Electronic Technology

基　　金：国家自然科学基金(61461010;61363031);认知无线电与信息处理省部共建教育部重点实验室基金(CRKL170103;CRKL170104);广西密码学与信息安全重点实验室基金(GCIS201626);广西高校云计算与复杂系统重点实验室基金(YF16203)

摘　　要：针对Android平台日趋增多的恶意软件安全问题,以及缺乏有效措施解决恶意软件的危害,提出一种基于Fisher归类算法结合本体推理的Android恶意软件安全检测方法。用Fisher分类算法对运行在手机端采集的应用程序所产生的系统调用频率信息进行处理,得出最佳线性判别函数,并据此求出最佳调用频率分界点,结合SWRL语言制定推理规则,应用本体JESS推理技术推理,实现了Android恶意软件的安全检测。实验结果表明,该方法可行、有效。For solving the issue of the increasing number of malware in Android and the lack of effective measurements,this paper put forward one new method based on Fisher classification algorithm and ontology inference to detect applications＇ security.First,the frequency of system calls were collected and the Fisher classification algorithm were used to process the frequency information,and the best linear discriminant function were concluded,then the best frequency dividing point of system calls were found according to the best linear discriminant function.Secondly,inference rules were defined in SWRL.Finally,with JESS inference technology,Android malware security detection was achieved.The experimental results proved its feasibility and validity.

关 键 词：安卓安全 SWRL规则 Fisher算法 JESS推理 本体 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]