Malware detection on the Android platform

OData support
Supervisor:
Dr. Buttyán Levente
Department of Networked Systems and Services

Nowadays, with the accelerating increase in the sale of mobile phones, and their important role in our everyday activities, they are perfect targets for malicious softwares. There is an exponential increase in the number of infected applications on the unofficial markets, and not even the official markets are safe. The Android platform is especially a good target because of it’s increasing popularity.

At first I analyze the current state of the Android malwares, and the existing detection methods. I examine in detail the variants of the DroidKungFu family, and the progress of their evolution.

In the second half of this paper, I examine the DroidKungFu variants in a virtual environment, and log their activities during their execution. Then I process, and organize these logs. After that I examine this organized dataset, using various methods provided by machine learning.

At last I evaluate my experiences with the different methods, and their usefulness in the detection of the chosen software.

Downloads

Please sign in to download the files of this thesis.