Topic Model Based Android Malware Detection

摘要: Nowadays, the security risks brought by Android malwares are increasing. Machine learning is considered as a potential solution for promoting performance of malware detection. For machine based detection, feature extraction plays key role. Thinking source codes applications comparable with text documents, we propose new detection method on topic model which an effective technique in extraction. Our regards decompiled application document, and used to mine topics can reflect semantic application. The experimental results demonstrate that, our approach performs better than state-of-the-art methods. Also, mines features files automatically without manually design, therefore overcomes limitation present methods relies experts’ prior knowledge.