Privacy Preserving Data Mining at Different Trust Levels

Abstract

Privacy preserving data mining has become
increasingly popular because it allows sharing of privacy sensitive
data for analysis purposes. So people have become increasingly
unwilling to share their data, frequently resulting in individuals either
refusing to share their data or providing incorrect data.The difficulty in
privacy-sensitive domain is solved by the development of the
Multi-Level Trust Privacy Preserving Data Mining (MLT-PPDM)
where multiple differently perturbed copies of the same data are
available to data miners at different trusted levels. In MLT-PPDM data
owners generate perturbed data by various techniques like Batch
generation and On-demand generation. MLT-PPDM can overcome the
diversity attacks. Partial information hiding methodologies like
random perturbation, random rotation perturbation are incorporated
with MLT-PPDM to enhance data security and to prevent leakage of
the sensitive data. The solution allows a data owner to generate
perturbed copies of its data for arbitrary trust levels on demand.
Finally MLT-PPDM approach is improved to tackle against the
non-linear attacks. The time and space complexities are calculated for
both techniques and the results show that on-demand algorithm is best
among them.