Detection and Classification of Malware Using Guided Whale Optimization Algorithm for Voting Ensemble

Marwa M. Eid^1*, M. I. Fath Allah²

¹ Faculty of Artificial Intelligence, Delta University for Science and Technology, Mansoura 35712, Egypt

² Communications and Electronics Department at Delta Higher Institute for Engineering and Technology, Mansoura- Egypt

Emails: marwa.3eeed@gmail.com; mismail1885@yahoo.com

Abstract

Malware is software that is designed to cause damage to computer systems. Locating malicious software is a crucial task in the cybersecurity industry. Malware authors and security experts are locked in a never-ending conflict. In order to combat modern malware, which often exhibits polymorphic behavior and a wide range of characteristics, novel countermeasures have had to be created. Here, we present a hybrid learning approach to malware detection and classification. In this scenario, we have merged the machine learning techniques of Random Forest and K-Nearest Neighbor Classifier to develop a hybrid learning model. We used current malware and an updated dataset of 10,000 examples of malicious and benign files, with 78 feature values and 6 different malware classes to deal with. We compared the model's results with those of current approaches after training it for both binary and multi-class classification. The suggested methodology may be utilized to create an anti-malware application that is capable of detecting malware on newly collected data.

Keywords: Cybersecurity; Malware detection; Machine learning; Hybrid learning; Classification; K-Nearest neighbor; Random forest; Metaheuristic optimization