Android平台下恶意软件分析与检测

Analysis and Detection of Malware Under the Android Platform

针对Android平台下恶意软件侵扰问题,提出一种基于权限—敏感API特征的加权朴素贝叶斯分类算法的检测方案。首先对Android应用程序中的配置文件进行解析,然后利用Apktool工具对APK文件进行反编译,提取出权限—敏感API特征集,并通过信息增益算法和卡方检验算法过滤冗余数据,最后利用加权朴素贝叶斯分类算法的恶意软件检测模型进行分类判断。实验结果证明,该系统能有效提高分类器的效率和恶意软件的检测率。

Aiming at the problem of malware intrusion under Android platform at present,this paper proposes an Android malware detection scheme based on the weighted naive Bayes classification algorithm based on the permission-sensitive API features.Firstly,the configuration files in the Android application is analyzed,Then the Apktool tool to decompile the APK file is used to extract the permission-sensitive API feature set,and the residual data in the feature set is filtered by the information gain algorithm and the Chi-square test algorithm;Finally,use the weighted naive Bayesian classification algorithm to maliciously The software detection model performs classification and judgment.The experimental results verify that the system can effectively improve the efficiency of the classifier and the detection rate of malware.