摘要
Android恶意软件对手机安全造成的威胁日益严重,为了防止恶意软件对手机造成的安全威胁,提出了一种从应用程序提取权限信息的方法,以此来检测软件的恶意行为。首先,在信息提取实验中,提取应用程序自身的权限信息和应用程序间具有权限提升威胁的信息,并对该信息进行统计分析,分别获得恶意软件和良性软件的差异和规律;其次,在检测实验中,根据提取实验的结果,利用机器学习和数据挖掘技术对应用软件进行分类,实现对恶意软件的静态检测。该方法能有效地对恶意软件进行预判断,达到检测的目的。实验结果表明,所提出的权限信息提取方法能较大地提高检测恶意软件的准确率。
The Android malware is becoming more and more serious security threat to mobile phone, so this paper proposed a method in order to prevent security threats from malware. First, it extracted the permission features in extraction experiment, which included the permission declare in each application and permission escalation between applications, analysed the difference between malware and benign with the extracted results. Then it used machine learning and data mining technology to detect the malware statically. The proposed method can effectively pre-judgment of the malware, and achieve the purpose of the detection. A conclusion can be drawn that the proposed method can greatly improve the accuracy of malicious detection.
出处
《计算机应用研究》
CSCD
北大核心
2015年第10期3036-3040,共5页
Application Research of Computers
基金
国家自然科学基金资助项目(61461010)
桂林电子科技大学研究生教育创新计划资助项目(GDYCSZ201413)
关键词
权限
恶意检测
安卓
机器学习
数据挖掘
permission
malware detection
Android
machine learning
data mining
