Detection and Analysis of Obfuscated File-less Malware Using Advanced Machine Learning Techniques

One of the biggest Internet concerns is malware. Now malware uses clever tactics to avoid antivirus protection. Modern cybersecurity protections have led to sophisticated and diverse file-less malware that runs without executable files. Fileless virus lives in memory without being detected by the file system. This unusual trait lets it bypass antivirus software and other security measures. Detecting file-less malware families and sub-families is improved in this study, improving incident response. Reducing detection time for each categorization stage (binary, family, subfamily) with machine learning algorithms and ensemble soft voting improves predicted accuracy. This method can improve file-less malware detection and subfamily classification, which aids mitigation. In binary classification, the proposed method achieves 99.9% accuracy, 88.86% in multiclass malware family classification, and 77.40% in subfamily classification. The evaluation findings show that the proposed strategy protects against complex malicious programs. When the specified characteristics were used, Binary classification detection time dropped from 3.359002 sec to (1.895786) sec, Family classification from 10.190352 to (3.694623) sec, and Sub-Family classification from (12.114225) to (3.737292) sec.