Detection of Android Malicious Family Based on Manifest Information

The current number of Android malicious applications is growing rapidly, which brings great troubles and losses to mobile phone users. Therefore, we propose a lightweight and efficient method for Android malicious family detection. The method uses DREBIN's Android malicious application dataset to build a detection model that using the application manifest information and machine learning algorithms. Through the training and testing of 10 types of Android malicious family samples, it is found that the detection model has the characteristics of low complexity and high classification accuracy, which can detect the Android malicious family very well and effectively protect the security of user mobile phones.