Abstract: In the contemporary digital era, the malware presents a substantial challenge to internet users, particularly with the rise of polymorphic malware that persistently alters its discernible features to evade detection by traditional signature-based models. This advanced malware variety, which boasts a dynamic design and inherits traits from multiple malware types, differs significantly from its conventional counterparts. The proposed research aims to examine and analyse the behaviour of malware executables, mainly focusing on their polymorphic attributes, to enhance cybersecurity through better understanding and detection.
Tackling the growing complexity and volume of malicious software is arduous, prompting researchers to employ machine learning techniques to decipher the underlying patterns and models within this intricate landscape, thereby keeping pace with malware's continuous evolution. This comprehensive review sheds light on using machine learning in the context of malware analysis for Windows environments, explicitly targeting Portable Executables.
In our analysis of the existing literature, we classify the studies based on their objectives, the malware-specific data and features, and the machine-learning approaches they apply. Furthermore, we delve into the challenges and issues tied to dataset usage and identify the dominant trends and promising future directions.
Keywords: Machine Learning, Cyber Security, Malware, Malware Analysis
|
DOI:
10.17148/IJARCCE.2023.12435
