作者: Huikang Hao , Zhoujun Li , Haibo Yu
DOI: 10.1109/TASE.2015.16
关键词:
摘要: As the most popular platform, Android dominates mobile device market. In order to enrich functions of phone and facilitate utilization users, more applications have been developed. Unfortunately, a greatly increasing amount malware targeting platform mingle with numerous benign hide in almost every market, even official market Google Play. Therefore, it is pressing concern about how measure assess risk such apps. this paper, we propose novel approach deal problem. First all, through empirical analysis market-scale dataset, verify following fact: for set same category, type number permissions they request are similar consistent general. Hence, each can construct standard permission vector model, which be used as baseline category. For downloaded app, extract its requested form vector, whose deviation from calculated by employing Euclidean distance weighted distance. The metric app. Finally, an experiment on real-world consisting 7737 apps 1260 samples, conducted evaluate our method. result validates effectiveness help users understand when decide install