Android Malware Detection Based on Runtime Behaviour

As the number of applications on Android markets grows, there is also a significant increase in the number of malicious applications that aim to harm users and devices. Therefore, mobile malware detection systems are developed and deployed for both Android markets and mobile devices. However, most malwares employ techniques such as code obfuscation, dynamic code loading in order to evade from static analysis based detection systems. For this reason, a dynamic analysis based detection method is proposed in this study. By examining the behaviour of malicious applications at runtime, features are extracted to distinguish them from benign applications, and a detection system is developed by using machine learning techniques.