A hybrid machine learning approach for analysis of stegomalware

Purpose Given how smart today’s malware authors have become through employing highly sophisticated techniques, it is only logical that methods be developed to combat the most potent threats, particularly where stealthy and makes indicators of compromise (IOC) difficult detect. After analysis completed, output can employed detect then counteract attack. The goal this work propose a machine learning approach improve detection by combining strengths both supervised unsupervised techniques. This study essential as has certainly ubiquitous cyber-criminals use attack systems in cyberspace. Malware required reveal hidden IOC, comprehend attacker’s severity damage find vulnerabilities within system. Design/methodology/approach research proposes hybrid for dynamic static combines algorithms goes on show exploiting steganography exposed. Findings tactics used developers circumvent are becoming more advanced with popular technique applied obfuscation evade mechanisms detection. continues call continuous improvement existing State-of-the-art approaches applying increasingly promising results. Originality/value Cyber security researchers globally grappling devising innovative strategies identify defend against threat extremely attacks key infrastructure containing sensitive data. process detecting presence requires expertise analysis. Applying intelligent aid practitioners identifying malware’s behaviour features. especially expedient stealthy, hiding IOC.