Microsoft-Malware-Challenge

The source code of feature extraction is not completely ready. But, The classification part is ready and you can run whole_system.py to see the results.

Confusion Matrix related to 5 Fold cross validation :


The final submitted file to kaggle website is in submissions folder.

The implementation is related to our paper at CODASPY'16:

https://www.researchgate.net/publication/283986464_Novel_Feature_Extraction_Selection_and_Fusion_for_Effective_Malware_Family_Classification

BibTex :

@inproceedings{Ahmadi:2016:NFE:2857705.2857713, author = {Ahmadi, Mansour and Ulyanov, Dmitry and Semenov, Stanislav and Trofimov, Mikhail and Giacinto, Giorgio}, title = {Novel Feature Extraction, Selection and Fusion for Effective Malware Family Classification}, booktitle = {Proceedings of the Sixth ACM on Conference on Data and Application Security and Privacy}, series = {CODASPY '16}, year = {2016}, isbn = {978-1-4503-3935-3}, location = {New Orleans, Louisiana, USA}, pages = {183--194}, numpages = {12}, url = {http://doi.acm.org/10.1145/2857705.2857713}, doi = {10.1145/2857705.2857713}, acmid = {2857713}, publisher = {ACM}, address = {New York, NY, USA}, keywords = {classification, computer security, machine learning, malware family, microsoft malware classification challenge, windows malware}, }

Requirements : numpy, pandas, xgboost, scikit-learn, hickle, pickle, numba, matplotlib

Thanks from Dmitry, Stanislav, and Mikhail for their support.