PIndroid: A novel Android malware detection system using ensemble learning

The extensive usage of smartphones has been the major driving force behind a drastic increase of new security threats. The stealthy techniques used by malware make them hard to detect with signature based intrusion detection and anti-malware methods. In this paper, we present PIndroid|a novel Permissions and Intents based framework for identifying Android malware apps. To the best of our knowledge, PIndroid is the first solution that uses a combination of permissions and intents supplemented with multiple stages of classifiers for malware detection. Ensemble techniques are applied for optimization of detection results. We apply the proposed approach on 1,745 real world applications and obtain 99.8% accuracy which is the best reported to date. Empirical results suggest that our proposed framework built on permissions and intents is effective in detecting malware applications.