Authors: Jie Wang, Jun Zhang, Justin Zhan
Addresses: Computer Information Systems, Indiana University Northwest, IN 46410, USA. ' Computer Science Department, University of Kentucky, Lexington, KY 40506-0027, USA. ' Heinz College & Cylab Japan, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA, 15213, USA
Abstract: Hiding data values in privacy-preserving data mining (PPDM) protects information against unauthorised attacks while maintaining analytical data properties. The most popular models are designed for constant data environments. They are usually computationally expensive for large data sizes and have poor real-time performance on frequent data growth. Considering that updates and growth of source data are becoming more and more popular in online environments, a PPDM model that has quick responses on the data updates in real-time is appealing. To increase the speed and response of the singular value decomposition (SVD) based model, we have applied an improved incremental SVD-updating algorithm. The performance and effectiveness of the improved algorithm have been examined on synthetic and real data sets. Experimental results indicate that the introduction of the incremental matrix decomposition produces a significant increase in speed for the SVD-based data value hiding method, better scalability and better real-time performance of the model, thereafter. It also provides potential support for the use of the SVD technique in the online analytical processing for business data analysis.
Keywords: data privacy; real-time performance; singular value decomposition; SVD; privacy preservation; data protection; privacy protection; data mining; modelling; online analytical processing; business data analysis.
International Journal of Granular Computing, Rough Sets and Intelligent Systems, 2010 Vol.1 No.4, pp.329 - 342
Available online: 20 Nov 2010 *Full-text access for editors Access for subscribers Purchase this article Comment on this article