A Privacy Protection System in Context-aware Environment The Privacy Controller Module

Abstract

As context-aware applications are becoming increasingly popular, there are also mounting demands for privacy protection systems. In our work, we propose a context-aware privacy protection system that consists of three modules and aims to recognize the user privacy behavior, classify the context-aware applications and recommend a set of protection action scenarios for the user privacy profile settings. Each module is a challenging problem that needs to be addressed using supervised and unsupervised Machine Learning (ML) algorithms. Part 1 of our work, this paper, consists of deploying hybrid techniques to handle the privacy controller module tasks. Logistic Regression (LR) learning algorithm is integrated with Statistical Method (SM) to recognize user privacy complex activities. The potential of the proposed system is demonstrated using a large-scale real-world dataset provided by institutes from Kuwait, the United States and Belgium. The system demonstration shows promising results with an accuracy of 97.9%.