Prevention for Direct and Indirect Discrimination in Data Mining
Data mining helps to extract useful and expected information among the huge amount of collective data present in database. Automated data collection with data mining collectively performs automated decisions. Discrimination can be direct or indirect. Direct discrimination use sensitive data for decision making. Indirect discrimination makes decisions on the basis of non-sensitive data. For more accuracy they express the relationship between discrimination prevention and privacy preservation in
... cy preservation in data mining. Along with security and privacy, proper discrimination performs vital role in considering legal as well as ethical point of view of data mining. The main aim behind this paper is to develop new preprocessing discrimination prevention methodology which consist of different types of data transformation methods. With the help of that direct discrimination, indirect discrimination or both of them at the same time get prevented. For making the final decision there are two steps in which first step include identification of categories and makes groups of individuals whatever it may be, directly indirectly discriminated for making decision. In second step with the help of clustering, transformation of data in specific way such that removes all discrimination.