Predicting Users Mobile App Privacy Preferences
Journal of Computer and Communications
Home users are using a wide and increasing range of different technologies, devices, platforms, applications and services every day. In parallel, home users are also installing and using an enormous number of apps, which collect and share a large amount of data. Users are also often unaware of what information apps collect about them, which is really valuable and sensitive for them. Therefore, users are becoming increasingly concerned about their personal information that is stored in these
... stored in these apps. While most mobile operating systems such as Android and iOS provide some privacy safeguards for users, it is unrealistic to manage and control a large volume of data. Accordingly, there is a need for a new technique, which has the ability to predict many of a user's mobile app privacy preferences. A major contribution of this work is to utilise different machine learning techniques for assigning users to the privacy profiles that most closely capture their privacy preferences. Applying privacy profiles as default settings for initial interfaces could significantly reduce the burden and frustration of the user. The result shows that it's possible to reduce the user's burden from 46 to 10 questions by achieving 86% accuracy, which indicates that it's possible to predict many of a user's mobile app privacy preferences by asking the user a small number of questions.