Distributed Mining of Association Rules Based on Privacy-Preserved Method

With the rapid development of social information, the application of distributed database system is increasing. Distributed data mining will play an important role in data mining. As one of the well-known distributed association rules mining algorithm, the FDM algorithm is very fast and efficient, however, the cost of this algorithm is very great because it is designed under the condition of non-shared resource. Moreover, the important information at every site is exposed to other sites, which is not accord to the nowadays trend of attaching importance to privacy preserving increasingly. In this paper, we propose an improved algorithm based on the FDM algorithm. In the process, it computes the total support count with the privacy-preserved method, meanwhile ensures the source of every local large item-set and local support count is covered, so it reduces the time spent on communication and preserves the privacy of the data distributed at each site. The experimental evaluations show that the proposed algorithm is efficient and rather suitable for the practical application field.