Classifying Socially Sensitive Data Without Discrimination: An Analysis of a Crime Suspect Dataset

Abstract

Abstract-Social discrimination against certain sensitive groups within society (e.g., females, blacks, minorities) is prohibited by law in many countries. To prevent discrimination arising from the use of discriminatory data, recent data mining research has focused on methods for making classifiers learned over discriminatory data discrimination-aware. Most of these methods have been tested on standard classification datasets that have been tweaked for discrimination analysis rather than over actual discriminatory data. In this paper, we study discrimination-aware classification when applied to a realworld dataset of Statistics Netherlands, which is a census body in the Netherlands. Specifically, we consider the use of classifiers for predicting whether an individual is a crime suspect, or not, to support law enforcement and security agencies' decision making. Our results show that discrimination does exist in real world datasets and blind use of classifiers learned over such datasets can exacerbate the discrimination problem. We demonstrate that discrimination-aware classification methods can mitigate the discriminatory effects and that they lead to rational and legally acceptable decisions.

Controlling Attribute Effect in Linear Regression

“…In this setting, a discrimination-ignorant regression model based on historical data can further exacerbate the racial discrimination. This has been demonstrated recently for crime suspect prediction using real-world data [14].…”

“…Thus, a discrimination-ignorant approach to wage prediction based on historical data (which is often biased) can lead to violation of laws. Similarly, racial discrimination in criminology and police arrests is a continuing concern [12]- [14]. In this setting, a discrimination-ignorant regression model based on historical data can further exacerbate the racial discrimination.…”

Ageism in the era of digital platforms

“…Thus, their algorithms are aligned with their objectives (Cheney-Lippold, 2011) and keeping them hidden helps to ensure that the corporate ideology remains invisible (Cheney-Lippold, 2011). In this sense, there is evidence of discrimination in credit approval decisions that take into account race, income, and zip code, among other aspects (Pedreschi et al, 2009); and in intelligent systems used to support the decision-making of security agencies (Kamiran et al, 2012). However, none of these studies consider ageism to be a significant source of discrimination.…”

Exploiting reject option in classification for social discrimination control

“…Our experience with the Dutch Research and Documentation Center (WODC) associated with the Ministry of Security and Justice and Statistics Netherlands, the national census body, confirms the importance of tackling discrimination w.r.t. multiple factors including age, gender, and race [18]. Being restricted to a specific discriminationaware classifier (e.g., naive Bayes [6], decision tree [17]) is also an issue because that classifier may not be the best performing classifier for a given dataset.…”