An Importance Analysis Of Android Applications' Permission Groups

In recent years,Android system has become the most popular mobile operating system,a large number of applications in the Android market provide convenient service for users.At the same time,Android applications use permissions to access user's privacy information,which could cause users' privacy leaking.Android officials also develop the Runtime Permissions System which proposes Permission Groups to package permissions and protect users' privacy information.However,because ordinary users lack understanding of permission groups and can't get the usage of permission groups in applications,it is difficult for them to make a reasonable settings for permission groups.Therefore,analyzing the importance of permission groups becomes an urgent problem to be solved.Although existing research proposes many analyze methods for permission problem of Android applications,there is no relevant research work for the importance of Permission Groups concept.We propose an importance analysis method for permission groups: First,use static analysis to analyze structure of program codes,mapping the codes and permission groups.Second,use pagerank algorithm to calculate importance score of permission groups,learning the usage and importance of Android applications.Finally,the experiment was designed to use data mining to judge the analysis results.We select the most accurate method to evaluate the importance of permission groups.We download 261 applications in the application market WANDOUJIA to verify our method.The experiment results show that our analysis method can analyze the importance of permission groups in detail and recommend the appropriate setting strategy for permission groups.This article contributes as follows:(1)As Android system changing its permission mechanism,we research usage on the new Permission Groups which proposed in the new mechanism(2)We use ranking algorithm to solve problems appearing in Android static analysis,sort importance of permission groups that used in the Android application,and recommend setups for permission groups.(3)We use machine learning to analyze permission groups used in Android applications,which provides a new application analysis method and a new classification idea.