Decoding the Secrets of Machine Learning in Windows Malware Classification: A Deep Dive into Datasets, Features, and Model Performance

This repository hosts the feature importance scores of the following experiments:

1. Binary detection
1. Static
1. All samples
2. Packed only
3. No packed
2. Dynamic
1. All samples
2. Family classification
1. Static
1. All samples
2. Packed only
3. No packed
2. Dynamic
1. All samples

Citation

If you use any of the contents, please cite it as:

@inproceedings{dambra2023decoding,
  title={Decoding the secrets of machine learning in malware classification: A deep dive into datasets, feature extraction, and model performance},
  author={Dambra, Savino and Han, Yufei and Aonzo, Simone and Kotzias, Platon and Vitale, Antonino and Caballero, Juan and Balzarotti, Davide and Bilge, Leyla},
  booktitle={Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communications Security},
  pages={60--74},
  year={2023}
}