/Z2F

Z2F: Heterogeneous Graph-Based Android Malware Detection

Primary LanguagePythonMIT LicenseMIT

Z2F

Z2F: Heterogeneous Graph-Based Android Malware Detection

Floder:decompile

Step1:using apktoo.jar and apktool.bat to install apktool to decompile Android packages; Step2:using decompile.py to batch decompile Android packages.

Floder:filter_samples

using check_empty.py,check_smali.py,delete_empty.py and filter_samples.py to check and delete all waste samples;

Floder:Android features extraction

After decompilation, we get a folder for each Android application; Step1:using extract_api.py to extract api feature; Step12:using extract_interface.py to extract interface feature; ...

Floder:features_data

We use JSON format to store the extracted raw data，api data not uploaded due to storage capacity；

file:hadmard_normalize.py

Refer to the paper 's original formula (1) ,formula (2);

model

The model is based on HAN

embedding

the samples's data after Z2F