SVD-based collaborative filtering with privacy

Collaborative filtering (CF) techniques are becoming increasingly popular with the evolution of the Internet. Such techniques recommend products to customers using similar users' preference data. The performance of CF systems degrades with increasing number of customers and products. To reduce the dimensionality of filtering databases and to improve the performance, Singular Value Decomposition (SVD) is applied for CF. Although filtering systems are widely used by E-commerce sites, they fail to protect users' privacy. Since many users might decide to give false information because of privacy concerns, collecting high quality data from customers is not an easy task. CF systems using these data might produce inaccurate recommendations. In this paper, we discuss SVD-based CF with privacy. To protect users' privacy while still providing recommendations with decent accuracy, we propose a randomized perturbation-based scheme.