An Android Malware Detection Method Based on Deep AutoEncoder
Nengqiang He, Tianqi Wang, Ping-Yang Chen, Hanbing Yan, Zhengping Jin · 2018
With the emergence of various Android malwares, many detection algorithms based on machine learning have been proposed to minimize their threat. However, those still have many shortcomings for detecting the emerging Android malware, thus some deep learning algorithms have already been applied to Android malware detection, but to the best of our knowledge deep AutoEncoder has not yet. In this paper, an Android malware detection method based on deep AutoEncoder is proposed, where a specify AutoEncoder structure is designed to reduce the dimension of feature vectors which are extracted and converted from APK, and the logistic regression model is also applied to learn and classify the Android applications to be normal or not. The experimental results show the recall rate and F1 value of our proposal can respectively reach 0.93 and 0.643, which perform better than other three similar models.