International Conference on Electronics, Information and Communication Engineering (EICE 2012)
116 Privacy Preserving Association Rules Mining Based on Concept Hierarchy
Download citation file:
Aimed at the privacy problems generated in the process of association rules mining, this paper proposes a security data mining method of single-layer association rules, and especially focuses on privacy preserving of data mining based on the data with concept hierarchy. First, the paper layered the information which was automatically generated by data statistics on the basis of conception hierarchy to realize cross-layer analysis. Second, to ensure security in the mining, a method was proposed which applied a secure invertible two-party matrix protocol to the multilayer association rules mining. Finally, by an example, the method was verified that can not only acquire the across-concept information, but also effectively protect the security of privacy.