Malware-Detection in android applications
About
- Permissions based Machine Learning model to detect whether an apk file is malware.
- Used Androguard tool to decompile apks. (https://github.com/androguard/androguard).
Usage
- Step-1 Clone this repository.
- Step-2 Open terminal in Malware-Detection directory.
- Step-3 Run the following command "pip install -U androguard"
- Step 4 Install all the required packages using command - "sudo pip3 install -r requirements.txt"
- Step-5 Run the follwing command - "python detector.apk path_to_apk"
Limitations
- Not trained on latest apk files.
- Only considers permissions in apk files to generate results.
Future work includes
- Use of latest dataset to train the model.
- More detailed hyperparameter tuning for the models
- Training Deep-Learning based model for better results.
Resources
-
Daniel Arp, Michael Spreitzenbarth, Malte Huebner, Hugo Gascon, and Konrad Rieck "Drebin: Efficient and Explainable Detection of Android Malware in Your Pocket", 21th Annual Network and Distributed System Security Symposium (NDSS), February 2014
-
Michael Spreitzenbarth, Florian Echtler, Thomas Schreck, Felix C. Freling, Johannes Hoffmann, "MobileSandbox: Looking Deeper into Android Applications", 28th International ACM Symposium on Applied Computing (SAC), March 2013
-
Research Paper - https://deepblue.lib.umich.edu/bitstream/handle/2027.42/136197/YouchaoDong_Thesis_0327.pdf%3Fsequence%3D1%26isAllowed%3Dy
Link for status presentation:- https://docs.google.com/presentation/d/1dsPaI0xw3dAgZQ3EJ3CKuuUWc_1gW4Sfl-GjwGEw_ng/edit#slide=id.p