“…Because we acquired statistical data through static analysis for the detection and classification of malware, this information was used as features in machine learning. Many features extracted by static methods, such as byte sequence [31], strings [31,43], DLL [31], n-gram [35], grayscale images [38], control flow graph (CFG) [39], function length frequency [40], PE header [41,42], mnemonics [49,50], API call [51][52][53][54], and opcode [8,[44][45][46][47], have typically been leveraged to detect and classify malware using machine learning. We select opcode as a core feature to distinguish malware from benign samples during execution.…”