Comprehensive Evaluation on Computational Intelligence Techniques for Malware Analysis

Malicious software has constituted threats to system security and operations carried out online due to the Internet and other technologies and associated breaches. In the past, scholarly research focused on the use of classical methods of malware detection and associated attributes like signature and heuristics, etc. Conventional malware detection methods become ineffective against emerging malware versions and their advanced evasion strategies. Lately, there have been attempts to put to use sophisticated Machine Learning (ML) for malware detection due to traditional techniques being highly ineffective on emerging variants of malware. This paper focuses on and implements methods for detecting and identifying the malware’s features using hash values associated with the Virus Total and the Quick Heal dataset. Also, this research surveys the recent ML techniques in malware detection and focuses on ML methods' sophistication and parameters like precision, recall, and the used datasets and methods. In addition, this research identifies in depth the highly understudied area of malware analysis.