A Simple Method for Detection of Metamorphic Malware using Dynamic Analysis and Text Mining

Metamorphic malware are the most challenging threat in digital world, which are quite advanced and have actually reduced the significance of signature based detection. These malware use code obfuscation to mutate and have numerous forms thus increasing the size of signature database; make it unmanageable and incomplete to cover all variants. This is the major reason why no anti-virus company can claim 100% detection even for non zero day malware.When a malware is encrypted or packed, Static analysis is not possible. In such cases, Dynamic analysis appears to be most obvious solution. But the challenge lies in finding out how to analyze behavior to detect malware in automated manner and how to quantize behavior. An approach is required here that specifies how to analyze dynamic report and how can we prepare a model which can help make detection decision. This is what we will be looking for in this paper.