Novel Android Malware Detection Method Based on Multi-dimensional Hybrid Features Extraction and Analysis

摘要

In order to prevent the spread of Android malware and protect privacy information from being compromised, this study proposes a novel multidimensional hybrid features extraction and analysis method for Android malware detection. This method is based primarily on a multidimensional hybrid features vector by extracting the information of permission requests, API calls, and runtime behaviors. The innovation of this study is to extract greater amounts of static and dynamic features information and combine them, that renders the features vector for training completer and more comprehensive. In addition, the feature selection algorithm is used to further optimize the extracted information to remove a number of extraneous features, and a new multi-dimensional hybrid features vector is obtained. The multi-dimensional hybrid features vector is then used to train the classification model. Finally, the unknown samples are detected and identified by using the obtained classification model. Our experiment is conducted based on 359 malicious and 500 benign applications as experimental samples, and the results indicate that our proposed method performs better in the accuracy rate of Android malware detection compared with those methods using static methods alone.