Benchmarking Open-Source Android Malware Detection Tools

As Android is becoming more and more prevalent in modern devices, malware targeting Android systems is significantly increasing. New generation of malware has great diversity and uses more sophisticated techniques to hide traces and fool traditional detection methods. On the other hand, many researchers have proposed various techniques and tools for detecting and classifying Android malware. The evaluation of most of these tools is based on closed or obsolete datasets, and likely the reported performance metrics are biased towards specific features or categories of malware. In this paper, we intensively evaluate the performance of an existing open-source tool, known as Dr-Droid, and benchmark its performance against several online standard malware detection engines freely accessible through VirusTotal website. The dataset has been recently revealed and covers a large number of Android malware categories and families.