Efficient machine learning approach for classification of malware with portable executable (PE) headers

As new forms of ransomware emerge every year, gadgets need to be restructured and secured. With that said, it is important to distinguish ransomware from other types of malware, as it occupies a unique category. Ransomware not only employs deceptive techniques but also performs actions such as locking or encrypting files that can severely damage systems. Since Signature-based detection algorithms are inherently unable to identify new zero-day malware, which underscores the need for more adaptive methods. Due to signatures-based forms of detection algorithms being fundamentally unable to identify new forms of zero-day malware, novel strategies are required. As such, Ransomware poses a unique threat which requires a whole new detection protocol that enables the identification of specific traits from malicious behaviors, mechanisms designed to hide from detection systems as well as other malware. This paper outlines a systematic approach to analyze the inner workings of malicious software tailored specifically for the reasons stated above.