Proposed Precautions for Newborn Malware Family Inspired from the COVID19 Epidemic Outbreak

Abstract

Malicious software attacks cause serious loss to computer users, from personal usage to industrial networks. For this reason, researchers focused more and more on analyzing and detecting malware. Approaches found in literature can well predict a new malware sample belonging to known families, but what about newborn families. In this paper, we perform malware classifiers based on two machine learning algorithms: Random forest and K-Nearest Neighbor. We used the malware visualization technique, so a malware binary is presented as a grayscale image. After that, we calculated the GIST descriptor features for all samples to be ready for the classification. Results obtained are, respectively, 97, 99%, and 98,21% for the two algorithms. Then, we study the behavior of our classifier in the case of new arrival family. Next, we get inspired from the COVID19 disease outbreak. So we proposed precautions to be made as security measures in case of new malware family appearance. With the goal to reduce damage causes by these kind of attacks.