PPAdroid: An Approach to Android Privacy Protocol Analysis

Abstract

With the continuous growth of the number of mobile applications, users may provide personal information to applications consciously or unconsciously. Privacy protocol could help users understand the privacy behavior of applications. However, users usually ignore the content of the privacy protocol due to the length of their context. To solve such a problem, we conduct an empirical study on privacy protocols of Android applications and propose a method to detect sentences related to personal information operations in privacy protocol documents. In our proposed method, a verb list, a negation list, and a noun list related to user information operations are created and utilized to detect sentences related to user personal information by the Stanford CoreNLP technique. The experimental results show that our proposed method is better than state-of-art keyword-based methods. Furthermore, our proposed method can help users understand the contents of privacy protocol documents in a limited way.