SVD-Based Factorization Technique for Dual Privacy Protection Data Mining

Abstract

Singular value decomposition (SVD) method is a very important matrix decomposition method in linear algebra. It is widely used in signal processing, statistics, data compression and other fields. The paper introduces a SVD method to reduce dimension of original dataset and makes use of the attribute of LSA technique to combine SVD method with LSA technique, and then presents new methods for dual private protection data mining. Finally we conduct experiments to test and verify the proposed approach and get good results.