Detection of Novelties in Malware Families

Authors

  • Ricardo Sant’Ana Instituto Militar de Engenharia
  • Julio Cesar Cardoso Tesolin Instituto Militar de Engenharia
  • Julio Cesar Duarte Instituto Militar de Engenharia

DOI:

https://doi.org/10.22491/rmct.v40i4.12122.pt

Keywords:

Gaussian Mixture Model, Malware Family Detection, Novelty Detection, Support Vector Machine, Malware as an image, Entropy

Abstract

Many researches have already presented approaches to the malware detection task. Classifying them into families provides a better understanding of their behavior, allowing companies and researchers to optimize their efforts. Nevertheless, an issue still needs to be proper addressed: how to verify if an artifact detected as a malware belongs to a known family? This work proposes the use of two widely known classifiers - GMM and SVM - for a novelty detection task in malware analysis to redirect proper human and computational efforts for a quick counter measure. The main contribution of this work is the use of features directly extracted from the detected malwares’s binary file such as entropy and image’s texture for novelty detection.

Downloads

Download data is not yet available.

Published

2026-04-06

How to Cite

Sant’Ana, R., Tesolin, J. C. C., & Duarte, J. C. (2026). Detection of Novelties in Malware Families. Revista Militar De Ciência E Tecnologia, 40(4). https://doi.org/10.22491/rmct.v40i4.12122.pt