Android Malware Detection Using Deep Learning
Omar N. Elayan, Ahmad Mustafa
Abstract
The Android operating system ranks first in the market share due to the system’s smooth handling and many other features that it provides to Android users, which has attracted cyber criminals. Traditional Android malware detection methods, such as signature-based methods or methods monitoring battery consumption, may fail to detect recent malware. Therefore, we present a novel method for detecting malware in Android applications using Gated Recurrent Unit (GRU), which is a type of Recurrent Neural Network (RNN). We extract two static features, namely, Application Programming Interface (API) calls and Permissions from Android applications. We train and test our approach using CICAndMal2017 dataset. The experimental results show that our deep learning method outperforms several methods with accuracy of 98.2%.