Improving Public Procurement Collusion Detection With Graph-based Machine Learning Methodologies

• Everton Schneider dos Santos UFSC
• Márcio Castro UFSC
• Jonata Tyska Carvalho UFSC

DOI: https://doi.org/10.5753/eramiars.2025.16657

Resumo

Public procurement is a complex process often susceptible to corruption and Machine Learning (ML) has emerged as a promising approach to identifying fraud in public procurement. While many ML methods for fraud detection rely on tabular data, information from the network of relationships between the entities involved in procurement process remains underutilized. This work aims to fill this gap with a study of ML methodologies for detecting collusion in public procurement using data extracted from the relationships network. Enriching the “Operation Car Wash” with topological information extracted from graphs helped improve the results by 1% and decrease the variability of the evaluated models by almost 5%.

Referências

Curtis, F. and et al. (1973). Closed competitive bidding. Omega, 1(5):613–619.

dos Santos, E. S., dos Santos, M. M., Castro, M., and Carvalho, J. T. (2024). Performance variability of machine learning models using limited data for collusion detection: A case study of the brazilian car wash operation. In Anais do XXXIX Simpósio Brasileiro de Bancos de Dados, pages 431–443, Porto Alegre, RS, Brasil. SBC.

Fazekas, M. and Wachs, J. (2020). Corruption and the network structure of public contracting markets across government change. Politics and Governance, 8(2):153–166.

Gallego, J., Rivero, G., and Martínez, J. (2021). Preventing rather than punishing: An early warning model of malfeasance in public procurement. International Journal of Forecasting, 37(1):360–377.

García Rodríguez, M. J. and et al. (2022). Collusion detection in public procurement auctions with machine learning algorithms. Automation in Construction, 133:104047.

Hamilton, W. L. (2020). Graph representation learning. Morgan & Claypool Publishers.

Pedregosa, F. and et al. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830.

Sanz, I. P., Iturriaga, F. J. L., and Blanco-Alcántara, D. (2024). A neural network approach for predicting corruption in public procurement. European Journal of International Management, 22(2):175–197.

Schneider dos Santos, E., Machado dos Santos, M., Castro, M., and Tyska Carvalho, J. (2025). Detection of fraud in public procurement using data-driven methods: a systematic mapping study. EPJ Data Science, 14(1).

Wallimann, H. and Sticher, S. (2023). On suspicious tracks: machine-learning based approaches to detect cartels in railway-infrastructure procurement. Transport Policy, 143:121–131.