The DOI matches its Crossref record; 8 of the 10 references it cites were found in a registry or an index.
Classifying without discriminating
Abstract
Classification models usually make predictions on the basis of training data. If the training data is biased towards certain groups or classes of objects, e.g., there is racial discrimination towards black people, the learned model will also show discriminatory behavior towards that particular community. This partial attitude of the learned model may lead to biased outcomes when labeling future unlabeled data objects. Often, however, impartial classification results are desired or even required by law for future data objects in spite of having biased training data. In this paper, we tackle this problem by introducing a new classification scheme for learning unbiased models on biased training data. Our method is based on massaging the dataset by making the least intrusive modifications which lead to an unbiased dataset. On this modified dataset we then learn a non-discriminating classifier. The proposed method has been implemented and experimental results on a credit approval dataset show promising results: in all experiments our method is able to reduce the prejudicial behavior for future classification significantly without loosing too much predictive accuracy.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Much of the previous work has primarily focused on settings where the outcome or protected variable is binary.”
From An algorithm for removing sensitive information: Application to race-independent recidivism prediction (2019) · cites this paper as Kamiran and Calders (2009)
In the citing paper
“Methodology for achieving statistical parity has centered on removing information about the protected variables from the training data by transforming the training data.”
From An algorithm for removing sensitive information: Application to race-independent recidivism prediction (2019) · cites this paper as Kamiran and Calders (2009)
In the citing paper
“This procedure is limited to binary outcome data and the adjusted data is not re-usable in the sense that one could not then use the covariates to estimate relationships with other outcome variables and still be ensured of nondiscriminatory outcomes.”
From An algorithm for removing sensitive information: Application to race-independent recidivism prediction (2019) · cites this paper as Kamiran and Calders (2009)
In the citing paper
“Prior work on detecting discrimination has focused on measuring correlation or mutual information between inputs and outputs [79], discrepancies in the fractions of inputs that produce a given output [18, 19, 39-42, 87-89, 91], or discrepancies in output probability distributions [51].”
From Fairness testing: testing software for discrimination (2017) · cites this paper as [39]
Citations
Cites
(1973) · CERN Document Server (European Organization for Nuclear Research) · Cited by 12,895
UCI Repository of machine learning databases
Blake, C. L. (1998) · Medical Entomology and Zoology · Cited by 10,582
Pattern Classification and Scene Analysis.
Clarke, M., Duda, R. O., & Hart, P. E. (1974) · Journal of the Royal Statistical Society Series A (General) · doi:10.2307/2344977 · Cited by 3,258
Samarati, P., & Sweeney, L. (2018) · doi:10.1184/r1/6625469 · Cited by 1,323
Discrimination-aware data mining
Pedreshi, D., Ruggieri, S., & Turini, F. (2008) · Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/1401890.1401959 · Cited by 676
Interestingness of frequent itemsets using Bayesian networks as background knowledge
Jaroszewicz, S., & Simovici, D. A. (2004) · ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/1014052.1014074 · Cited by 128
Fast discovery of unexpected patterns in data, relative to a Bayesian network
Jaroszewicz, S., & Scheffer, T. (2005) · doi:10.1145/1081870.1081887 · Cited by 44
Knowledge-Based Sampling for Subgroup Discovery
Scholz, M. (2005) · Lecture notes in computer science · doi:10.1007/11504245_11 · Cited by 12
Cited by
Sorted by citations
Items 401–403 of 403
Exploring Backdoor Vulnerabilities and Fairness Challenges in Generative Models: A Survey
Holland, R., Pal, S., & Pan, L. (2026) · ACM Computing Surveys · doi:10.1145/3847107
The Impact of Bias Mitigation on Fairness and Accuracy in Automated Skin Lesion Classification
Rocha, M. B., & Krohling, R. A. (2026) · Journal of Healthcare Informatics Research · doi:10.1007/s41666-026-00252-w
Integrated Tool for Evaluating Ethical Principles in AI: An Extensive Case Study
Teixeira, S., Cortés, Á., Thilakarathne, D., Gori, G., Minici, M., Bhuyan, M., Khairova, N., Adewumi, T., Bhuyan, D., O’keefe, J., et al. (2026) · Springer eBooks · doi:10.1007/978-3-031-61050-9_39-1
Metadata
- Authors
- Kamiran, F., & Calders, T.
- Affiliations
- Eindhoven University of Technology (NL)
- Venue
- 2009 2nd International Conference on Computer, Control and Communication
- Publisher
- IEEE
- Published
- 2009-02
- Pages
- 1-6
- OpenAlex
- W2157928966
- Cited by
- 409
- References
- 10
- Citations by year
- 2026: 15 · 2025: 26 · 2024: 43 · 2023: 57 · 2022: 48 · 2021: 52 · 2020: 51 · 2019: 46 · 2018: 27 · 2017: 11 · 2016: 5 · 2015: 7 · 2014: 3 · 2013: 1 · 2012: 8
- FWCI
- 28.77
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- International Conference on Computer, Control and Communication (publishedVersion)
- TU/e Research Portal (publishedVersion)
- TU/e Research Portal (submittedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Imbalanced Data Classification Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Keywords
- algorithmic discrimination, biased training data, racial discrimination, credit approval, algorithmic fairness, prediction accuracy
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
