An efficient protocol for private and accurate mining of support counts

  • Authors:
  • Fan Wu;Jiqiang Liu;Sheng Zhong

  • Affiliations:
  • Department of Computer Science and Engineering, State University of New York at Buffalo, Buffalo, NY 14260, USA;Department of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China;Department of Computer Science and Engineering, State University of New York at Buffalo, Buffalo, NY 14260, USA

  • Venue:
  • Pattern Recognition Letters
  • Year:
  • 2009

Quantified Score

Hi-index 0.10

Visualization

Abstract

In recent years, a large number of data mining tools were developed, which may reveal costumers' privacy if proper protection measure is not taken. On the other hand, customers are becoming increasingly concerned about privacy. They are reluctant to provide personal information unless privacy-preserving techniques are used. In this paper, we propose a privacy-preserving protocol for mining support counts, which maintains high accuracy and strong privacy while achieving very good efficiency. Compared with existing works with similar privacy and accuracy guarantees, our solution is much more efficient. We use identity-based cryptography, which has an additional advantage that no public key certificate is needed. Further, our evaluation results show that the protocol is very efficient and practical.