基于CNN和朴素贝叶斯方法的安卓恶意应用检测算法

Android Malware Detection Algorithm Based on CNN and Naive Bayesian Method

安卓系统已经成为移动端市场占有率领先的操作系统,但是安卓市场上恶意应用泛滥的情况十分严重,这些恶意应用会给用户带来极大的威胁.如何提前检测安卓应用程序是否为恶意应用程序已成为亟待解决的问题.提出了一种检测算法,该算法使用操作码序列和权限信息作为应用的特征,并分别用基于卷积神经网络(convolution neural network, CNN)和朴素贝叶斯方法的分类器进行分类,从而达到检测安卓恶意应用的目的.结果表明提出的算法效果稍好.

Android has become the leading operating system in the mobile market. However, the flooding of malicious applications on the Android market is very serious. These malicious applications will bring users great threats. How to detect whether an Android application is a malicious application in advance has become an urgent problem to be solved. In this paper, our scheme improves the original algorithm and proposes a method, which uses both the opcode sequence and the permission information as the characteristics of application and classifies them by convolution neural network (CNN) and naive Bayesian method respectively. This method can detect in advance whether the Android application is a malicious application. The results show that our method works slightly better.