A Novel Framework for Metamorphic Malware Detection

Malwares are a major threat in the evolving global cyberspace. The different techniques for anti-virus software, in which presently there is insufficiency in detecting metamorphic malwares as they can change their internal structure of the code, keeping the flow of the program equivalent to the virus. Commercial Antivirus software depends on signature detection algorithms to identify viruses, but code obfuscation techniques can circumvent the above algorithms successfully. The objective of this research is to analyze the different detection techniques of such metamorphic malware. We also propose a novel methodology of detecting them via use of different machine learning algorithms, such as KNN, Support Vector Machine (SVM), RF (random forest), and naive Bayes. We also establish multiple semantic preserving transformation techniques for code obfuscation. Analysis regarding the same has been presented too. © 2022, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.