Mining Frequent Patterns Securely in Distributed System

Jiahong WANG  Takuya FUKASAWA  Shintaro URABE  Toyoo TAKATA  Masatoshi MIYAZAKI  

Publication
IEICE TRANSACTIONS on Information and Systems   Vol.E89-D   No.11   pp.2739-2747
Publication Date: 2006/11/01
Online ISSN: 1745-1361
DOI: 10.1093/ietisy/e89-d.11.2739
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Data Mining
Keyword: 
data mining,  distributed computing,  privacy-preserving,  performance evaluation,  

Full Text: PDF(514KB)>>
Buy this Article




Summary: 
Data mining across different companies, organizations, online shops, or the likes is necessary so as to discover valuable shared patterns, associations, trends, or dependencies in their shared data. Privacy, however, is a concern. In many situations it is required that data mining should be conducted without any privacy being violated. In response to this requirement, in this paper we propose an effective distributed privacy-preserving data mining approach called SDDM. SDDM is characterized by its ability to resist collusion. Unless the number of colluding sites in a distributed system is larger than or equal to 4, privacy cannot be violated. Results of performance study demonstrated the effectiveness of SDDM.