Implementation Of Privacy Preservation Using Anonymization Methods For Discrimination Prevention
Privacy preserving data mining (PPDM) refers to the part of data mining used to safeguard sensitive information illegal disclosure. Discrimination is the detrimental process of people based on their association with a certain classes or groups. Direct discrimination restricts a certain group of working class based on sensitive reasons. Indirect discrimination restricts a certain group of working class based on non sensitive ones. Both direct and indirect discrimination can be prevented using
... prevented using data transformation methods such as rule protection and rule generalization. Balanced iterative reducing and clustering using hierarchies (BIRCH) algorithm is used for analyzing discrimination datasets based on eligible criteria. In this paper, privacy can be enhanced using differentiated virtual password schemes and anonymization techniques. We provide a differentiated virtual password that applies user-specified randomized linear generation functions to protect user passwords. We provide an anonymization algorithm that processes inferring approach to prevent attacks in discrimination environment. We are evaluating these methods on Adult dataset and provide metrics for proposed methods that impact on information loss and data quality in data mining.