The vast majority of modern mobile malware targets Android devices. This made it necessary to investigate efforts in the analysis of malicious Android programs in recent years. An important task of malware analysis is the classification of malware.
This article presents the work in which are presented in the new methodology classify malicious Android software on their own families. The proposed approach is based on the dynamic analysis and use of indicators. This approach to classify malware with an accuracy of 82%. This is the result of the DroidScribe, the most modern work that is most connected with ours. However, this methodology has the following advantages:
– it is easier to reproduce, because only uses publicly available tools and does not require any modification of environment emulation or Android OS;
– the collected data can be collected on a physical device.
In the future, the authors of the plan to increase the number of indicators that can be achieved in the study. And also there are prerequisites for the sake of learning this online learning to reduce the time needed to update the classifier and, therefore, timely classify new families of malware.
