Privacy Preserving Data Mining Using Additive Perturbation on Relational Streaming Data

Data mining concerns with extracting the required important data from the database and ignoring the rest. With the success of data mining, privacy preservation has also acquired the great importance. The new concept privacy preserving data mining PPDM, concerns with preserving the privacy of sensitive individuals data. In this paper, privacy of sensitive attribute data concerned with individual user or data miner is preserved. For preserving the privacy additive perturbation method is used, in which random noise are added to the sensitive attribute values from the require data set and perturb copies are generated. A new concept multilevel trust MLT-PPDM approach is used, in which we generate multiple perturb copies of same data for the data miners at different trust level. For perturb copies generation, group generation algorithm is used, in which for a given original data, multiple perturbed copies of same data will be generated. We are using relational streaming database which means records in the database are updated continuously and at the same time for each updated records perturb copies will be generated successfully, which is the new contribution to the proposed work.