Deep Analysis and Utilization of Malware’s Social Relation Network for Its Detection

摘要: To combat with the evolving malware attacks, many research efforts have been conducted on developing intelligent detection systems. In most of existing systems, resting analysis file contents extracted from samples (e.g., binary n-grams, system calls), data mining techniques such as classification and clustering used for detection. However, ignoring social relations among these (i.e., utilizing only) is a significant limitation methods. this paper, (1) instead using collected samples, we conduct deep relation network study how it can be detection; (2) constructed graph, perform large scale inference by propagating information labeled (either benign or malicious) to detect newly unknown malware. A comprehensive experimental collection sample obtained Comodo Cloud Security Center performed compare various approaches. Promising results demonstrate that accuracy efficiency our proposed method outperform other alternate based techniques.