Rule Protection for Indirect Discrimination Prevention in Data Mining
- Cite this paper as:
- Hajian S., Domingo-Ferrer J., Martínez-Ballesté A. (2011) Rule Protection for Indirect Discrimination Prevention in Data Mining. In: Torra V., Narakawa Y., Yin J., Long J. (eds) Modeling Decision for Artificial Intelligence. MDAI 2011. Lecture Notes in Computer Science, vol 6820. Springer, Berlin, Heidelberg
Services in the information society allow automatically and routinely collecting large amounts of data. Those data are often used to train classification rules in view of making automated decisions, like loan granting/denial, insurance premium computation, etc. If the training datasets are biased in what regards sensitive attributes like gender, race, religion, etc., discriminatory decisions may ensue. Direct discrimination occurs when decisions are made based on biased sensitive attributes. Indirect discrimination occurs when decisions are made based on non-sensitive attributes which are strongly correlated with biased sensitive attributes. This paper discusses how to clean training datasets and outsourced datasets in such a way that legitimate classification rules can still be extracted but indirectly discriminating rules cannot.
KeywordsAnti-discrimination Indirect discrimination Discrimination prevention Data mining Privacy
Unable to display preview. Download preview PDF.