Feature Analysis and Risk Assessment of Android Group Based on Clustering

Abstract

The security risk assessment of Android applications is an uncertain problem. It is difficult to determine whether the application authority is reasonable from a single application point of view, and whether some privacy rights overstep the functional scope of the application itself. To solve this problem, a group based method for feature analysis and risk assessment of large-scale Android applications is proposed. The permission that app applies is an important object for security analysis and evaluation. The same type of application has similar functions, so the required system privileges are similar. By comparing the application class of the same functional type to the population, the two layers model of group feature analysis and mass population clustering is used to evaluate the relative malicious program in the population, which shows the effectiveness and adaptability of the method.