Computing scalar products amongst private vectors in a secure manner is a frequent operation in privacy-preserving data mining algorithms, especially when data is vertically partitioned on many parties. Existing secure scalar product protocols based on cryptography are costly, particularly when they are performed repeatedly in privacy-preserving data mining algorithms. To address this issue, we propose an efficient cacheable secure scalar product protocol called CSSP that is built upon a homomorphic multiplicative cryptosystem. CSSP allows one to reuse the already cached data and thus, it greatly reduces the running time of any privacy-preserving data mining algorithms that adopt it. We also conduct experiments on real-life datasets to show the efficiency of the protocol. © 2011 Springer-Verlag.
CITATION STYLE
Tran, D. H., Ng, W. K., Lim, H. W., & Nguyen, H. L. (2011). An efficient cacheable secure scalar product protocol for privacy-preserving data mining. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6862 LNCS, pp. 354–366). https://doi.org/10.1007/978-3-642-23544-3_27
Mendeley helps you to discover research relevant for your work.