Enabling Multi-level trust in privacy Preserving Data mining

By using the aspect of Multi-level Preservation of privacy in Data mining, accurate modules about aggregate data can be built without accessing to the information in individual copy of the data record. Based on the multilevel perturbation - PPDM approach generates privacy before the data is being published by the perturbation approach. Existing approach for the problem is Privacy in individual security level on the database(db's). So we are going to provide ease to this assumption and expand the scope of perturbation-based PPDM to Multilevel Trust (MLT-PPDM). In our scenario, the data miner should be more specialized, so it does not access more distributed copy of the module. Under this scenario, the fraud data miner by using different aspects can access to distributed copies of the data, and may aggregate these different copies together to get precise information about the data which data owner would never want to get leaked. Preventing and maintaining such diversity attacks is the key challenge of providing MLT-PPDM services. We will address this challenge by properly correlating perturbation across copies at different trust levels. We will provide a solution which will be robust against diversity attacks by aiming to our privacy goal. So our solution will facilitate attacker who have access to an different data set of the distributed copies, our solution prevent them from rebuilding the data which is actual and more accurately than using any individual copy in the database(db's). We provide owner of data to generate distributed copies of its data for multi-level trust (MLT-PPDM) on demand. This feature gives maximum flexibility and intractability to data owners.