Physical Sciences and Mathematics Commons^™

Malware Detection Using Dynamic Analysis, Swapna Vemparala

Master's Projects

In this research, we explore the field of dynamic analysis which has shown promis- ing results in the field of malware detection. Here, we extract dynamic software birth- marks during malware execution and apply machine learning based detection tech- niques to the resulting feature set. Specifically, we consider Hidden Markov Models and Profile Hidden Markov Models. To determine the effectiveness of this dynamic analysis approach, we compare our detection results to the results obtained by using static analysis. We show that in some cases, significantly stronger results can be obtained using our dynamic approach.