Acquiring and Analyzing App Metrics for Effective Mobile Malware Detection

摘要: Android malware is becoming very effective in evading detection techniques, and traditional techniques are demonstrating their weaknesses. Signature based shows at least two drawbacks: first, the possible only after has been identified, time needed to produce distribute signature provides attackers with window of opportunities for spreading wild. For solving this problem, different approaches that try characterize malicious behavior through invoked system API calls emerged. Unfortunately, several evasion have proven evade on calls.In paper, we propose an approach capturing terms device resource consumption (using a thorough set features), which much more difficult camouflage. We describe procedure, corresponding practical setting, extracting those features aim maximizing discriminative power. Finally, promising results obtained experimenting than 2000 applications, our exhibited accuracy greater 99%.