Fractal methods for the representation and analysis of polymorphism in malware

The preponderance of network resident cyber threats are due to zero day vulnerabilities or unpatched systems. Traditional signature based detection methods are ineffective against such threats, and anomaly detection methods are typically computationally intensive. By treating polymorphism as a stochastic process and applying fractal visualization methods, identifying features can be found. These features are suited to then be used for detection with algorithms such as Bayes classifiers.