Program for Malware Apk-Files Detection Based on Machine Learning Algorithms

Android OS has become the most popular mobile operating system for mobile phones and tablets. These devices are used to store gigabytes of sensitive personal information such as photos, contacts, emails credit and banking information. Even part of this data might be very important and that is why smartphones might be a victim of hackers. Gaining access to a device, perpetrators might perform some unwanted actions – steal personal data including banking information, show advertisement, download other applications and many other potential malicious actions. The most common way to gain access to a device and use it for bad intents is to install APK file. APK files available from the online markets, which often do not filter the content and may contain millions of malicious APK files. The number of new malware APK is increasing every day and detect it manually is impossible and some automated system is needed. This paper describes a machine learning approach for malicious APK files detection based on permissions and calls of Android OS API within the APK file. The designed system has acceptable detection and false positive rates.