The Open Automation and Control Systems Journal

2015, 7 : 800-804
Published online 2015 August 19. DOI: 10.2174/1874444301507010800
Publisher ID: TOAUTOCJ-7-800

Research on a New Privacy-Preserving Algorithm of Association Rules Based on Parameter Perturbation

Jingjing Yang , Beibei Dong , Xiao Zhang , Zhonghua Li and Shangfu Hao
School of Information Science and Engineering, Hebei North University, Zhang Jiakou, Hebei, 075000, China.

ABSTRACT

Due to the reason that the randomness of the parameters in the MASK algorithm always leads to the volatility and uncertainty of the mining results, this paper proposed an optimization algorithm for the maximum likelihood estimation of the parameters to choose a parameter that is most approximate to the common parameters from the parameter group that has been generated randomly. Such a parameter generated as above represented all of the parameters in the parameter group. The simulation experiment proves that the application of such a parameter has reduced the great volatility hidden in the mining results to some extent.

Keywords:

Association rules, Privacy preserving, Parameter optimization.