基于SVM的敏感权限Android恶意软件检测方法

SVM Method of Detecting Android Malicious Software Based on Sensitive Permissions

Android系统使用权限机制来控制其应用程序访问系统资源和用户的隐私信息,因此,Android应用程序所申请的权限与其行为关系紧密,通过分析权限信息可以有效地检测Android恶意软件。本文提出了一种基于敏感权限的Android恶意软件SVM检测方法,通过恶意软件逆向分析,统计分析两库中各个权限的出现频率和恶意软件库中的敏感权限,构建恶意软件与非恶意软件权限特征库。最后,通过SVM方法建立检测模型有效地对恶意软件进行判断。

Android system uses permission mechanism to control the application access to system resources and users＇ a private information,therefore,the permissions applied for by Android application programs is closely related to their behaviors. Through the analysis of authority information can effectively detect Android malicious software. This paper proposes a kind of SVM method of detecting Android malicious software. By malicious software reverse analysis,statistical analysis of the frequency of each authority in the two banks and the sensitive permissions of malicious software,we are to construct software access feature database of malicious software and non-malicious software. Finally,by using the SVM method,we are to establish the detection model to judge the malicious software effectively.