The Research Of Association Rules Mining Based On Privacy Preserving Techniques

Recently, with the improvement of capability of data storage, the problems of privacy preserving comes to catch our attention. The subject of data mining, as a powerful knowledge-finding tool, can perfectly extract useful information from a sea of data. At the same time, it lets a host of privacy data out with the process of data mining. Data mining with privacy preserving had leapfrog development in a few years. For all kinds of technology of data mining, it provides the methods of privacy preserving accordingly. From the experiment data, those algorithms present wonderful result, while, it has some defects in practical applicability aspect which makes that privacy preserving data mining technology has a profound study prospect.This article mainly discusses the method of privacy preserving for association rule mining in the distributed environment. The main work is as follows:1. Describe data mining and association rule mining method comprehensively, introduce the concept of privacy disclosure and classify it. Classify and summarize the privacy preserving methods for association rule algorithms;2. For the privacy preserving of source data in the distributed environment, we implement privacy preserving method for mining of association rule using the technology of homomorphic encryption. And we improve the algorithm by introducing the technology of data disturbance which works properly without leaking privacy under the circumstance of malevolence involvement. At the same time, consideration of illegal involvement, we provide a validation method of zero-knowledge proving to serve for practical distributed application.3. We precede a deep research on the algorithm of hiding sensitive rules and suggest a Border-based Approach. The aim of the algorithm is to reduce the lose of non-sensitive rules caused by sensitive rules hidden as much as possible.4. At the stage of publishment of association rule, we study quantization method and introduce several important algorithms.5. Design a perfect platform of mining of association rules based on privacy preserving, by integrating relative technology using the content of research.