Malicious Android Application Detection Based on Composite Features

Abstract

With the use of mobile phones, malicious applications are constantly developing, affecting the normal use of mobile phones by users. For the malicious application of Android platform, a detection model based on combined features is proposed. The model extracts the dynamic and static features and select the importance of them. Selecting Combination Features from important features. Taking the combined features as new features, and combing the single features to detect Android malicious applications. Experiments are carried out using different classification algorithm. which verifies the proposed Android malicious application detection model is feasible and superior, and the detection accuracy is up to 97.12%.