Inferring Property Typology in Real Estate Tax Records via Domain-Aware Feature Alignment

• Luís Henrique Alves Rosa UFG
• Gustavo Luiz Bueno Pereira UFG
• Lucas Araújo Pereira UFG
• Luiz Mário Lustosa Pascoal UFG
• Sávio Salvarino Teles de Oliveira UFG

DOI: https://doi.org/10.5753/eniac.2026.31554

Resumo

Real estate tax records in Brazil typically contain transaction values and built areas but lack structural attributes such as bedroom counts, limiting comparative market analyses. We propose a source-domain feature engineering approach to infer bedroom counts in unlabelled tax records using labelled listing portal data. To address the domain shift caused by the gap between announced and declared prices (66% median difference), we introduce an intra-neighbourhood relative price metric that corrects the shift without cross-dataset parameter estimation. Evaluated on a held-out listing test set, k-Nearest Neighbours (k=15) achieves 93.4% accuracy, while a density-based pipeline (UMAP and HDBSCAN) achieves 80.9%, providing interpretable clusters and per-record uncertainty estimates (ECE = 0.026). A zero-shot cross-portal experiment on 6,882 external listings yields 84.4% accuracy, validating the portability of the relative price feature. The method is evaluated on 1,542 tax transactions across 39 neighbourhoods in Goiânia, Brazil.

Palavras-chave: ITBI, domain shift, property typology, UMAP, HDBSCAN, KNN, urban informatics

Referências

Afonso, B. K. D. A. (2019). Housing prices prediction with a deep learning and random forest ensemble. In Anais do XVI Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), Porto Alegre, RS, Brasil. SBC.

Akiba, T., Sano, S., Yanase, T., Ohta, T., and Koyama, M. (2019). Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 2623–2631.

Alves, E. L. G., Viana, L. Z. B., Santos, F., Franco, W., and Pinheiro, V. (2023). Leveraging property tax collection through an unsupervised model for taxpayer qualification. In Anais do XX Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), Porto Alegre, RS, Brasil. SBC.

Antipov, E. and Pokryshevskaya, E. (2012). Mass appraisal of residential apartments: an application of Random Forest for valuation and a CART-based approach for model diagnostics. Expert Systems with Applications, 39(2):2772–2778.

Azevedo, A. C. and Santos Filho, R. C. d. (2025). Predictive modeling of real estate prices using machine learning. In Anais do XXII Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), Porto Alegre, RS, Brasil. SBC.

Ben-David, S., Blitzer, J., Crammer, K., Kulesza, A., Pereira, F., and Vaughan, J. (2010). A theory of learning from different domains. Machine Learning, 79(1–2):151–175.

Brasil (1988). Constituição da república federativa do Brasil de 1988, Art. 156, Inciso II. Diário Oficial da União.

Campello, R., Moulavi, D., and Sander, J. (2013). Density-based clustering based on hierarchical density estimates. In Advances in Knowledge Discovery and Data Mining, volume 7819 of LNAI, pages 160–172, Heidelberg. Springer.

Carneiro, A. and Centeno, J. (2003). Atualização cadastral multifinalitária utilizando imagens de alta resolução e SIG. Revista Brasileira de Cartografia, 55(1):1–10.

Carvalho Jr, P. H. B. d. (2014). Distributive aspects of real estate property and its taxation among brazilian families. Texto para discussão, Instituto de Pesquisa Econômica Aplicada (IPEA), Brasília, Brasil.

Courty, N., Flamary, R., Tuia, D., and Rakotomamonjy, A. (2017). Optimal transport for domain adaptation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(9):1853–1865.

Elmagarmid, A., Ipeirotis, P., and Verykios, V. (2007a). Duplicate record detection: A survey. IEEE Transactions on Knowledge and Data Engineering, 19(1):1–16.

Elmagarmid, A. K., Ipeirotis, P. G., and Verykios, V. S. (2007b). Duplicate record detection: A survey. IEEE Transactions on Knowledge and Data Engineering, 19(1):1–16.

Ganin, Y. and Lempitsky, V. (2015). Unsupervised domain adaptation by backpropagation. In Proceedings of the 32nd International Conference on Machine Learning (ICML), pages 1180–1189.

Guo, C., Pleiss, G., Sun, Y., and Weinberger, K. (2017). On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning (ICML), pages 1321–1330. PMLR.

Levantesi, S. and Piscopo, G. (2020). The importance of interest rates in building residential real estate price predictions with machine learning: the case of Italy. Risks, 8(3):92.

Long, M., Cao, Y., Wang, J., and Jordan, M. (2015). Learning transferable features with deep adaptation networks. In Proceedings of the 32nd International Conference on Machine Learning (ICML), pages 97–105.

Malpezzi, S. (2003). Hedonic pricing models: a selective and applied review. Housing Economics and Public Policy, pages 67–89.

McInnes, L., Healy, J., and Melville, J. (2018). UMAP: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426.

Pan, S. and Yang, Q. (2010). A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10):1345–1359.

Quiñonero-Candela, J., Sugiyama, M., Schwaighofer, A., and Lawrence, N., editors (2009). Dataset Shift in Machine Learning. MIT Press, Cambridge.

Sirmans, G. S., Macpherson, D. A., and Zietz, E. N. (2005). The composition of hedonic pricing models. Journal of Real Estate Literature, 13(1):1–44.

You, Q., Pang, R., Cao, L., and Luo, J. (2017). Image-based appraisal of real estate properties. IEEE Transactions on Multimedia, 19(12):2751–2759.