Aiming at the security problems of Android mobile application,the malicious application detection method was deeply studied,and a malicious Android application detection method based on privilege features was proposed.In the method,an algorithm to mine the frequent itemsets of rights was designed,which was called DroidFP-Growth.When constructing the permission relational feature library,this algorithm was used to mine the frequent itemsets of the sample sets to obtain the detection rules.This algorithm only need to scan the sample set twice to obtain the frequent itemsets, which effectively improved the efficiency of the construction of the authority relation database and improved the accuracy of the detection.The final results showed that the method of detection rate of malicious applications reached 81.2%,the accuracy rate of 83.6%,compared with similar methods also had certain advantages.%针对Android手机应用程序存在的安全问题,对恶意应用的检测方法进行了深入研究,提出一种基于权限特征的Android恶意应用检测方法.方法中设计了一种挖掘权限频繁项集的算法——DroidFP-Growth.在构建权限关系特征库时,利用该算法挖掘样本集的权限频繁项集,获得检测规则.该算法仅需扫描两次样本集便可获得权限频繁项集,有效地提高了构建权限关系特征库的效率,同时也提高了检测的准确率.最终实验结果表明,方法对恶意应用的检测率达到81.2%,准确率达到83.6%,对比同类方法也一定优势.
展开▼
