Privacy aware recommender service for IPTV networks

Providers of the next generation of IPTV services seek to gain competitive advantage over competing providers. In order to attract and satisfy customers, these providers must offer added value e.g. by delivering suitable content according to customers personal interests in a seamless way. This can achieved using recommender services. However, this brings about additional requirement related to the privacy of users' profiles that has to be addressed to make these services widely accepted. The ability to deploy a privacy aware recommender services and in the same time provide accurate recommendations become a key success for the spread of these services. Nevertheless, Current implementations of recommender services are mostly centralized where the users' profiles are stored in single server. This implementation fails to achieve the required privacy guarantee for the users, as it requires collecting accurate private information. In this paper, we present our efforts to build a private centralized recommender service (PRS) using collaborative filtering techniques by introducing an agent based middleware (AMPR) to ensure user profile privacy in the recommendation process. The driving force for using software agents in this work is the autonomic intelligent behaviour that can be achieved using agent technology. AMPR preserve the privacy of its users when using the system and allow private sharing of data among different user in the network. We also introduce two-stage obfuscation process embedded in AMPR that protect user profile privacy and preserve the aggregates in the dataset to maximize the usability of information in order to get accurate recommendations. These processes give the user a complete control on the privacy level of his personal profile. We also provide an IPTV network scenario and experimentation results.