Analysis on Malware Detection with Multi Classifiers on M0Droid and DroidScreening Datasets

Abstract

The number of applications for smart mobile devices is steadily growing with the continuous increase in the utilization of these devices. the Installation of malicious applications on smart devices often arises the security vulnerabilities such as seizure of personal information or the use of smart devices in accordance with different purposes by cyber criminals. Therefore, the number of studies in order to identify malware for mobile platforms has increased in recent years. In this study, permissionbased model is used to detect the malicious applications on Android which is one of the most widely used mobile operating system. M0Droid and DroidScreening data sets have been analyzed using the Android application package files and permission-based features extracted from these files. In our work, permission-based model which applied previously across different data sets investigated to M0Droid and DroidScreening datasets and the experimental results has been expanded. While obtaining results, feature set analyzed using different classification techniques. The results show that permission-based model is successful on M0Droid and DroidScreening data sets and Random Forests outperforms another method. When compared to M0Droid system model, it is obtained much bet ter conclusions depend on success rate. Our approach provides a method for automated static code analysis and malware detection with high accuracy and reduces smartphone malware analysis time.