Abstract: Aiming at the problem of privacy information security in mobile applications, the author proposed a security analysis method for mobile application based on user interface (UI) content. Firstly, this method  got the functions of mobile applications by mining the information of UI  products, and determined sensitive permissions that the application actually used by analyzing the code. Secondly,  the applications with similar functions were clustered together by using Mean shift algorithm. Finally, based on principle that products with similar functions should  use similar sensitive permissions, the anomaly detection algorithm iForest was used to determine whether the product was at risk of use.  Experimental results show that this method can effectively analyze the security of mobile applications.

Key words: mobile application, safety, sensitive permission, user interface mining, anomaly detection