NADM: Neural Network for Android Detection Malware

Abstract

Over recent years, Android is always captured roughly 80% of the worldwide smartphone volume. Due to its popularity and open characteristic, the Android OS is becoming the system platform most targeted from mobile malware. They can cause a lot of damage on Android devices such as data loss or sabotage of hardware. According to the predictive characteristics, machine learning is a good approach to deal with the number of new malwares increasing rapidly. In this paper, we propose Neural Network for Android Detection of Malware (NADM). The NADM performs an analysis process to gather features of Android applications. Then, these data will be converted into joint vector spaces, which to be input for the training part of deep learning process. Our classifier model can achieve a high accuracy system and has been applied in sProtect [15] on Google Play.