Fairness in Risk Estimation of Brazilian Public Contracts

  • Órion Darshan Winter De Lima Universidade Federal de Campina Grande
  • Nazareno Andrade Universidade Federal de Campina Grande


Brazilian government agencies are currently using machine learning models to make public contracts audition through risk estimation. Recent works have shown that decision making models, like risk estimation, may be unfair. Despite the fact that risk estimations of public contracts may be unfair, no studies evaluating model fairness have been found. This work contributes by analysing fairness over risk estimation of brazilian public contract. This article found that currently used models are unfair and biased towards a specific class. This means that people within this class may be negatively affected by these decision making models unfairness through risk estimation of their companies.

Palavras-chave: fairness, machine learning, risk estimation


Lei 8666. In Constituição Federal, 1993.

Cumming, G. and Calin-Jageman, R. Introduction to the New Statistics: Estimation, Open Science, and Beyond.
Routledge, New York, NY, 10001, 2016.

Gomes, T. A., Carvalho, R. N., and Carvalho, R. S. Identifying anomalies in parliamentary expenditures of
brazilian chamber of deputies with deep autoencoders. In IEEE ICMLA, 2017.

Julia Angwin, Jeff Larson, S. M. and Kirchner, L. Machine bias: There’s software used across the country to
predict future criminals. and it’s biased against blacks., 2016.

Saleiro, P., Benedict Kuester, A. S., Anisfeld, A., Hinkson, L., London, J., and Ghani, R. Aequitas: A bias
and fairness audit toolkit. In eprint arXiv:1811.05577, 2018.

Saul M. Kassin, I. E. and Kukucka, J. The forensic confirmation bias: Problems, perspectives, and proposed
solutions. In Journal of Applied Research in Memory and Cognition. Vol. 2. pp. 42–52, 2013.

Silvio L. Domingos, Rommel N. Carvalho, R. S. C. and Ramos, G. N. Identifying it purchases anomalies in the
brazilian government procurement system using deep learning. In 15th IEEE ICMLA. pp. 722–727, 2016.

Speicher, T., Heidari, H., Grgic-Hlaca, N., Gummadi, K. P., Singla, A., Weller, A., and Zafar, M. B. A
unified approach to quantifying algorithmic unfairness: Measuring individual group unfairness via inequality indices. In ACM KDD ’18, 2018.

Sun, T. and Sales, L. J. Predicting public procurement irregularity: An application of neural networks. In Journal
of Emerging Technologies in Accounting: Spring 2018. Vol. 15. pp. 141–154, 2018.

Zan Huang, Hsinchun Chen, C.-J. H. W.-H. C. S. W. Credit rating analysis with support vector machines and
neural networks: a market comparative study. In Decision Support System. Vol. 37. pp. 543–558, 2004.
Como Citar

Selecione um Formato
DE LIMA, Órion Darshan Winter; ANDRADE, Nazareno. Fairness in Risk Estimation of Brazilian Public Contracts. In: SYMPOSIUM ON KNOWLEDGE DISCOVERY, MINING AND LEARNING (KDMILE) , 2019, Fortaleza. Anais do VII Symposium on Knowledge Discovery, Mining and Learning. Porto Alegre: Sociedade Brasileira de Computação, nov. 2019 . p. 57-64. DOI: https://doi.org/10.5753/kdmile.2019.8789.