Android Malware Detection Based on Multi-Features

With the widespread use of the Android system, the number of malicious Android applications increases sharply. How to effectively identify malware and improve the successful detection of malicious code becomes increasingly important. Traditionally, the detection technologies mainly focuse on the analysis of a single feature, and can not fully utilize the role of multiple types of features on Android malicious code. In this article, we first study the difference between the resource features of malicious applications and normal applications, and then propose an approach to detecting Android malicious code by combining permissions and API features with resource features. The experiments show that the accuracy of this method is better than the methods of using the permissions or API features alone, it will effectively compensate for the insufficiency of the traditional detection methods, and thus can be used effectively to detect Android malicious applications.