An Android Malware Detection Method Using Deep Learning based on Multi-features

As we all know, android is an operating system used by all walks of life, and many developers have developed a large number of applications. Because of its openness, it is becoming the target of malware, which may pose a serious security threat to all aspects of life. In order to deal with these security threats, we propose an excellent malware detection method based on text classification. This method uses androguard tool to extract permission, service, receiver and intent from android application package, and combines them into a text report representing the android application. We apply BiLSTM network to mine important information in the text. The accuracy of our experiment is 97.47%, so our system can effectively detect Android malware.