Privacy-preserving Similarity Sorting in Multi-party Model

Abstract

In social network, it is conceivable that a rational execution sequence does good to cooperative mission, especially for a large number of participants. However, there are many difficulties for multi-party computation, the most important of which is privacy. In this paper, secure multiparty computation technology and dimensionality reduction are chosen to design a privacy-preserving protocol, which sorts m people according to their similarity. In a n dimensional system, the secure protocol’s time complexity is O(mn+n+m logm) and communication complexity is O(m). Detailed analysis about security and applicability are also presented in this paper. In addition, the protocol can be improved in security at the cost of complexity, with an arbitration agreement designed against fraud.

Topics

    10 Figures and Tables

    Download Full PDF Version (Non-Commercial Use)