A Novel Feature Representation for Malware Classification. (arXiv:2210.09580v1 [cs.CR])

In this study we have presented a novel feature representation for malicious
programs that can be used for malware classification. We have shown how to
construct the features in a bottom-up approach, and analyzed the overlap of
malicious and benign programs in terms of their components. We have shown that
our method of analysis offers an increase in feature resolution that is
descriptive of data movement in comparison to tf-idf features.

classification malware malware classification representation