Review Paper on Prevention of Direct and Indirect Discrimination

Data mining is very important technology for extracting useful knowledge from large data. The discrimination is nothing but the unfair treatment given to an individual or group according to particular characteristics. For data mining classification rules are performing very important role but discrimination comes into picture because of biased classification rules. The training data sets are biased so we need to firstly discover discrimination and then need to prevent that discrimination to make it discrimination free. Discrimination can be of two types, direct and indirect. When decisions are made based on sensitive attributes, Direct Discrimination occurs. While decisions based on non-sensitive attributes, Indirect Discrimination occurs. The experimental evaluations demonstrate that the proposed techniques are effective at removing direct and/or indirect discrimination in the original data set while preserving data quality