Files
Download Full Text (1.2 MB)
Abstract
The rapid growth of the smartphone applications market raises security concerns regarding untrusted applications. Studies have shown that most apps in markets request to collect data irrelevant to the main functions of the apps. Traditional Android permission control design based on one-time decisions on installation has been proven to be ineffective in protecting users’ privacy and poorly utilizes scarce mobile resources (e.g. battery). RecDroid is designed to help protect user data in a market where unauthorized data collection is prevalent in most applications. With RecDroid’s recommendation feature, which is based on expert analysis, the user will be able to avoid unnecessary privacy breach while retaining safe functionalities of many third party applications. RecDroid’s implementation consists of three distinct components: Application, OS modifications, and Recommendation Server. Firstly, RecDroid is exposed to the user as an application that allows them to manage permissions for installed applications and provides recommended permissions for new installations. In order to accomplish this, the Android OS needed to be modified to add two key functions to make our application work: editing the PackageManager class to handle user set permissions and to add a probation installation setting to let RecDroid know what applications to monitor. The last core component required for the RecDroid application is a server that collects participating users’ decisions sent by the application, makes recommendations based on stored decisions, and provides recommendations to the application. When RecDroid is fully implemented and has a sufficient user base, we expect to have a fully functional application that can effectively protect user data from dishonest applications that seek to overstep their essential reach. The server will continuously collect more data and recruit new users, which will correspondingly increase the confidence level and accuracy of RecDroid recommendations.
Publication Date
2015
Keywords
computer science, recommendation service
Disciplines
Computer Engineering | Engineering
Faculty Advisor/Mentor
Carol Fung
VCU Capstone Design Expo Posters
Rights
© The Author(s)
Date of Submission
July 2015