An Efficient Privacy-Preserving Friend Recommendation Scheme for Social Network

Friend recommendation is important and inevitable in social network. However, many recommendation strategies of social networks are not necessarily consistent with user's interests. In order to avoid the randomness and unreliability of friend recommendation in social network, this paper proposed a safe and efficient friend recommendation scheme based on privacy-preserving protocol. Target users send the request for friends to Center Authority (CA), CA selects friend candidates from all users in system with tag matching on the basis of privacy-preserving protocol. Target users choose friends among friend candidates filtered by CA. What is more, the proposed algorithm selected an appropriate similarity degree value as the threshold through the similarity degree comparison experiments, and the similarities among keywords and users are well defined and evaluated.