Recently, privacy issues have become important in
data analysis, especially when data is horizontally
partitioned over several parties. In data mining, the
data is typically represented as attribute-vectors and,
for many applications, the scalar (dot) product is one
of the fundamental operations that is repeatedly used.
In privacy-preserving data mining, data is distributed
across several parties. The efficiency of secure
scalar products is important, not only because
they can cause overhead in communication cost, but
dot product operations also serve as one of the basic
building blocks for many other secure protocols.
Although several solutions exist in the relevant literature
for this problem, the need for more efficient
and more practical solutions still remains. In this
paper, we present a very efficient and very practical
secure scalar product protocol. We compare it to the
most common scalar product protocols. We not only
show that our protocol is much more efficient than the
existing ones, we also provide experimental results by
using a real life dataset. |
| Cite as: Amirbekyan, A. and Estivill-Castro, V. (2007). A New Efficient Privacy-Preserving Scalar Product Protocol. In Proc. Sixth Australasian Data Mining Conference (AusDM 2007), Gold Coast, Australia. CRPIT, 70. Christen, P., Kennedy, P. J., Li, J., Kolyshkina, I. and Williams, G. J., Eds. ACS. 209-214. |
 (from crpit.com)
(local if available)
|