Binary Information Press
To solve the problem of privacy disclosure in cross-system collaborative filtering recommendation, the authors propose a secure computation model based on the theory of RSA public-key cryptography and secure multi-party computation. The model can prevent the third parties from colluding to the systems. Applied it to the cross system recommendation, an effective privacy preserving collaborative filtering recommendation algorithm is presented. The algorithm uses scalar product protocol to compute the users' similarity across system, thus the users' ratings cannot be divulged.