Deep Learning Techniques for Android Botnet Detection

摘要: Android is increasingly being targeted by malware since it has become the most popular mobile operating system worldwide. Evasive families, such as Chamois, designed to turn devices into bots that form part of a larger botnet are becoming prevalent. This calls for more effective methods detection botnets. Recently, deep learning gained attention machine based approach enhance detection. However, studies extensively investigate efficacy various models currently lacking. Hence, in this paper we present comparative study techniques using 6802 applications consisting 1929 from ISCX dataset. We evaluate performance several including: CNN, DNN, LSTM, GRU, CNN-LSTM, and CNN-GRU 342 static features derived applications. In our experiments, achieved state-of-the-art results on dataset also outperformed classical classifiers.