Skip to content
PegXPublic

About

HP-MDC is an Android Malware Detection (one of the classic pattern recognition issues) framework. It combines malware’s patterns from network traffic and code graph structure.

Resources

Stars

2 stars

Watchers

1 watching

Forks

Latest commit

 

History

7 Commits

Folders and files

Repository files navigation

HP-MDC - Hybrid Pattern Malware Detection and Categorization

HP-MDC is an Android Malware Detection (one of the classic pattern recognition issues) framework. It combines malware’s patterns from network traffic and code graph structure.

AndroNetMnist Dataset

For the network traffic feature based module, we split the raw PCAP file to the network flow (groups network packet with a 5-tuple, the network flow should have the same (sourceID, sourcePort, destinationIP, destinationPort, Protocol)). And then we convert each network packet to a 784-byte gray image. In our case, we refer to the Android Adware and General Malware Dataset (CIC-AAGM2017) https://www.unb.ca/cic/datasets/android-adware.html. We created and released our dataset AndroNetMnist, which includes 3,255,391 2D gray images in five classes for network traffic classification. Please check out it with this link https://drive.google.com/drive/folders/1XkVuy2BuvbMKabELEUpBZ1sGK1iQWBmW?usp=sharing.

If you want to generate the AndroNetMnist dataset by yourself, please check the network_traffic folder nad README file under it.

Static feature from Binary and Graph Neural Network

For the program code feature based module, we introduce natural language processing inspired methods in our malware detection and categorization framework. The detailed information could be found in code_graph_embedding folder

About

HP-MDC is an Android Malware Detection (one of the classic pattern recognition issues) framework. It combines malware’s patterns from network traffic and code graph structure.

Resources

Stars

2 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages