Polymorphic malware detection using topological feature extraction with data mining

摘要: In just a few short years, the number of polymorphic and metamorphic malware samples seen in wild has grown exponentially, automated detection apparatus which is largely signature-based finds itself virtually practically useless for these new types attacks. New methods are needed order to better defend networks, protect data preserve overall internet operations. This paper offers novel approach extract, analyze combine multiple high level factors determine “malicious or benign” files. These include file characteristics, internal properties dynamic run-time relationships. presents unique leveraging topological examination using static analysis. Belief Propagation (BP) achieved through mining techniques uncover spotlight malicious The proposed directly captures file-properties can therefore identify files with impressive rates (.9999) low false positives (.0001). should prove be faster than large reputation database performs well small sample sizes.