Faculty of Informatics, Kogakuin University 1-24-2, Nishi-Shinjuku, Shinjuku, Tokyo, 163-8677 Japan.

---

Abstract—Many works have been done for privacy-preserving matching protocols. Most of them obtain one-to-one privacy-preserving matching. However, when we consider forming a group of people or objects by their similarity, matching can be applied to problems using many data, such as recommendation systems and a lot of similar communities. In this paper, we consider the characteristics of each user as a vector. We obtain the similarity by securely computing the inner product of vectors. Also, we define a group’s characteristics by the members’ average characteristics. We propose a privacy-preserving group matching protocol. We show computation cost of the proposed protocol and show simulation results.

Index Terms—Privacy-preserving computation, group matching, homomorpic encryption.

[PDF]

Cite: Takuya Ibaraki, Yoshifumi Manabe, "Privacy-Preserving Group Matching Protocol," Journal of Computers vol. 13, no. 9, pp. 1037-1041, 2018.