Detection of Conflicts between APP's Privacy Policy and Actual Behavior: A Security Analysis System

Leaks of user privacy in the mobile cloud environment have been common in recent years. Common APP stores require apps to have a user privacy policy that complies with applicable laws. Due to the problem of lengthy papers or difficult-to-understand sentences in privacy policies, users frequently skip reading them or fail to comprehend them. Moreover, there may be conflicts between the privacy policy and the actual behaviors. In order to alleviate these problems above, we design a security analysis system that employs natural language processing to detect conflicts between an APP's privacy policy and its actual behaviors. Experimental results show that the analysis accuracy is improved compared with existing methods.