Enabling multi level trust in privacy preserving data mining

To propose an additive perturbation based PPDM to address the problem of developing accurate models about all data without knowing exact information of individual values. To preserve privacy, the approach introduces random perturbation to individual values, before the data are published to third parties for mining purposes. In Existing System, the PPDM approach assumes single level trust on data miners. Under the single level trust, a data owner generates only one perturbed copy of its data with affixed amount of uncertainty. In proposed system, the PPDM approach introduces multilevel trust on data miners. Here different perturbed copies of same data are available to data miner at different trust levels & may combine these copies to jointly add additional information about original data & release the data is called diversity attacks. To prevent these attacks, using multilevel PPDM approach where random Gaussian noise is added to the original data with arbitrary distribution. So, the data miners will have no diversity gain in their joint reconstruction of the original data. This allows data owners to generate perturbed copies of its data on demand at arbitrary trust levels. It provides data owner very flexibility.