Detecting Targeted Smartphone Malware with Behavior-Triggering Stochastic Models

摘要: Malware for current smartphone platforms is becoming increasingly sophisticated. The presence of advanced networking and sensing functions in the device giving rise to a new generation targeted malware characterized by more situational awareness, which decisions are made on basis factors such as location, user profile, or other apps. This complicates behavioral detection, analyst must reproduce very specific activation conditions order trigger malicious payloads. In this paper, we propose system that addresses problem relying stochastic models usage context events derived from real traces. By incorporating particularities given user, our scheme provides solution detecting targeting user. Our results show properties these follow power-law distribution: fact facilitates an efficient automatic testing patterns tailored individual users, when done conjunction with cloud infrastructure supporting cloning parallel testing. We report empirical various representative case studies, demonstrating effectiveness approach detect complex patterns.