Onamd: An Online Android Malware Detection Approach

Abstract

Android mobile phone is one of the most usable smartphone today which makes Android the most anticipated operating system in the market. However, the increase in malware threats with the growth in Android market cannot be neglected. Android is an open source platform which gives full leverage to developers on one side, but also provides an open door to malware. Dynamic analysis provides a comprehensive view of malware detection, but it is still subjective to the high-cost development and manual endeavors in test analysis. In this article, we propose a mechanism based on static and dynamic features for malware detection in Android. We leverage the proposed technique and develop a system, called an online Android Malware Detection Approach ONAMD. Initially, the ONAMD extracts the information (e.g., requested permissions, and basic data info, etc Next, it applies SVM and Random Forest algorithm that enhances the malware modeling capability to classify the application as benign or malicious. We applied our approach to 600 applications. The experimental result shows the efficiency of our approach which takes halftime and better recalls rate than Androguard.