Malware Detection Using Machine Learning Based on the Combination of Dynamic and Static Features

Abstract

As millions of new malware samples emerge every day, traditional malware detection techniques are no longer adequate. Static analysis methods, such as file signature, fail to detect unknown programs. Dynamic analysis methods have low efficiency and high false positive rate. We need a detection technique that can adapt to the rapidly changing malware ecosystem. The paper presented a new malware detection method using machine learning based on the combination of dynamic and static features. The characteristic of this experiment involved in many fields of knowledge, including binary program instrumentation, static analysis, assembly instruction analysis, machine learning, etc. Finally, we achieved a good result over a substantial number of malwares.