Title :
A New Privacy-Preserving Support Vector Regression Model on Vertically Partitioned Data
Author_Institution :
Coll. of Inf. & Eng., Qingdao Binhai Univ. Qingdao, Qingdao, China
Abstract :
A novel model for privacy-preserving support vector regression (PPSVR for short) on vertically partitioned data is proposed in the paper. The feasibility of the model is proved. Besides, the algorithm for vertically partitioned data is given out. In the privacy preserving data mining, each entity is unwilling to share its group of data or leak the data for various reasons. The proposed PPSVR model is public. But no private data is revealed. And when the PPSVR is calculated at last, the original data does not need to be recovered. Besides, the proposed algorithm has comparable accuracy with that of an ordinary SVR that uses the centralized data set directly. Experiments show that the proposed approach is effective.
Keywords :
data mining; data privacy; regression analysis; support vector machines; PPSVR; PPSVR model; privacy preserving data mining; privacy-preserving support vector regression model; private data; vertically partitioned data; Data models; Data privacy; Partitioning algorithms; Security; Support vector machines; Vectors; Privacy-preserving; Support Vector Regression; vertically partitioned data;
Conference_Titel :
Intelligent Systems Design and Engineering Applications (ISDEA), 2014 Fifth International Conference on
Conference_Location :
Hunan
Print_ISBN :
978-1-4799-4262-6
DOI :
10.1109/ISDEA.2014.19
