作者: Younghee Park , Douglas S. Reeves , Mark Stamp
DOI: 10.1016/J.COSE.2013.09.006
关键词:
摘要: Detection of malicious software (malware) continues to be a problem as hackers devise new ways evade available methods. The proliferation malware and variants requires advanced methods detect them. This paper proposes method construct common behavioral graph representing the execution behavior family instances. generates one by clustering set individual graphs, which represent kernel objects their attributes based on system call traces. resulting has path, called HotPath, is observed in all instances same family. proposed shows high detection rates false positive close 0%. derived highly scalable regardless added. It also robust against attacks.