Three naive Bayes approaches for discrimination-free classification

被引:399
作者
Calders, Toon [1 ]
Verwer, Sicco [1 ]
机构
[1] Eindhoven Univ Technol, NL-5600 MB Eindhoven, Netherlands
关键词
Discrimination-aware classification; Naive Bayes; Expectation maximization;
D O I
10.1007/s10618-010-0190-x
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this paper, we investigate how to modify the naive Bayes classifier in order to perform classification that is restricted to be independent with respect to a given sensitive attribute. Such independency restrictions occur naturally when the decision process leading to the labels in the data-set was biased; e.g., due to gender or racial discrimination. This setting is motivated by many cases in which there exist laws that disallow a decision that is partly based on discrimination. Naive application of machine learning techniques would result in huge fines for companies. We present three approaches for making the naive Bayes classifier discrimination-free: (i) modifying the probability of the decision being positive, (ii) training one model for every sensitive attribute value and balancing them, and (iii) adding a latent variable to the Bayesian model that represents the unbiased label and optimizing the model parameters for likelihood using expectation maximization. We present experiments for the three approaches on both artificial and real-life data.
引用
收藏
页码:277 / 292
页数:16
相关论文
共 12 条
  • [1] CALDERS T, 2010, CONSTRUCTING DECISIO
  • [2] CALDERS T, 2009, IEEE ICDM WORKSH DOM
  • [3] Chan P. K., 1998, Proceedings Fourth International Conference on Knowledge Discovery and Data Mining, P164
  • [4] Duivesteijn W, 2008, LECT NOTES ARTIF INT, V5211, P301, DOI 10.1007/978-3-540-87479-9_38
  • [5] Elkan C., 2001, The Foundations of Cost-Sensitive Learning
  • [6] Kamiran F, 2010, P BEN
  • [7] KAMIRAN F, 2009, P IC409, pIC409
  • [8] KOTLOWSKI W, 2007, P ECML PKDD 07
  • [9] Margineantu D.D., 1999, LEARNING DECISION TR
  • [10] NIJSSEN S, 2007, P ACM SIGKDD