An Effective Performance Based on Static and Dynamic Features to Detect Malware in Android by Machine Learning Algorithm

Abstract

Android is dominating the smartphone market with more users than any other mobile operating system. Yet concern from hackers has also risen with its growing popularity, as the number of malicious applications remains to rise. Wide-ranging work on malware identification and prevention for Android devices has been carried out in recent years, although Android has also introduced numerous security measures to fix malware complications, containing Unique User ID (UID) for every request, software permissions, and its Google Play scattering platform. In our proposed work, incorporates the analysis of static and dynamic features of these applications with the goal of evaluating their actions by exploring different attributes such as authorization, use of CPUs and storage utilization. In this paper machine learning methods to evaluate the comparative efficiency of extracted static and dynamic features to identify AM. This results is quite powerful and demonstrates the AUC of 0.972 can be used to identify AM with a high degree of accuracy than the dynamic function.