Abstract: Data mining is the useful technology for extracting knowledge buried in large storage of data. Discrimination is one of the negative effects of data mining. Automatic collection of data and data mining techniques such as classification have cleared the way to make automatic decisions, like loan granting/denial, insurance premium computation, giving job etc. In classification, if the banking training datasets are discriminated towards any sensitive attributes like gender, age, religion, etc., then the resulting decisions will have discrimination like loan rejection. Discrimination discovery and discrimination prevention are the two anti-discrimination techniques introduced in data mining. Gender discrimination, age discrimination, racial discrimination are some of the examples of discrimination. But all these discriminations come under the two class namely, direct and indirect discrimination. If the resulting decision depends directly on sensitive attributes, it comes under direct discrimination. If it depends on background knowledge like census data, other than the sensitive attributes it is indirect. The discrimination problems in banking services like loan grant/deny are resolved with the help of anti-discrimination technique. Also the measures of discrimination are identified. The training data related to banking services are transformed in the correct way to remove the discriminatory biases, while maintaining data quality.
Keywords: Anti-discrimination, Classification, Pre-processing, Rule protection, Rule generalization.
Title: APPLYING ANTI-DISCRIMINATION TECHNIQUES IN BANKING SERVICES
Author: P. SWARNA REKHA, M. KAYATHRI DEVI
International Journal of Computer Science and Information Technology Research
ISSN 2348-120X (online), ISSN 2348-1196 (print)
Research Publish Journals
