Evaluating Computational and Carbon Costs of Machine Learning: The Impact of Feature Dimensionality During Model Development
Resumo
The expansion of AI increases energy demand, making Green AI crucial. This study evaluates computational and carbon trade-offs in financial credit risk across six architectures under feature variations (±10% to ±50%). Using CodeCarbon, we propose ECAR to quantify emissions per unit of predictive performance. Results show that across several model–dataset configurations, reducing feature dimensionality by 20% to 50% lowered emissions without significant losses in F1-score. In contrast, the evaluated feature augmentation increased computational costs without proportional gains in predictive performance. These findings highlight feature dimensionality and model architecture as key factors in the environmental efficiency of financial ML systems.Referências
Chen, T., He, T., Benesty, M., Khotilovich, V., Tang, Y., Cho, H., Chen, K., Mitchell, R., Cano, I., Zhou, T., et al. (2015). Xgboost: extreme gradient boosting. R package version 0.4-2, 1(4):1–4.
de Chavannes, L. H. P., Kongsbak, M. G. K., Rantzau, T., and Derczynski, L. (2021). Hyperparameter power impact in transformer language model training. In Proceedings of the second workshop on simple and efficient natural language processing, pages 96–118.
Delanoë, P., Tchuente, D., and Colin, G. (2023). Method and evaluations of the effective gain of artificial intelligence models for reducing co2 emissions. Journal of Environmental Management, 331:117261.
Georgiou, S., Kechagia, M., Sharma, T., Sarro, F., and Zou, Y. (2022). Green ai: Do deep learning frameworks have different costs? In Proceedings of the 44th International Conference on Software Engineering, pages 1082–1094.
Guo, G., Wang, H., Bell, D., Bi, Y., and Greer, K. (2003). Knn model-based approach in classification. In OTM Confederated International Conferences” On the Move to Meaningful Internet Systems”, pages 986–996. Springer.
Haakman, M., Cruz, L., Huijgens, H., and Van Deursen, A. (2021). Ai lifecycle models need to be revised: An exploratory study in fintech. Empirical Software Engineering, 26(5):95.
Jiao, Z., Wang, H., Xing, J., Yang, Q., Yang, M., Zhou, Y., and Zhao, J. (2023). A light-gbm based framework for lithium-ion battery remaining useful life prediction under driving conditions. IEEE Transactions on Industrial Informatics.
Kaur, R., Kundu, T., Park, K. M., and Pinsky, E. (2026). The carbon cost of intelligence: A domain-specific framework for measuring ai energy and emissions. Energies, 19(3):642.
Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., and Liu, T.-Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems, 30.
LaValley, M. P. (2008). Logistic regression. Circulation, 117(18):2395–2399.
Li, G., Yu, Z., Yang, K., Lin, M., and Chen, C. P. (2024). Exploring feature selection with limited labels: A comprehensive survey of semi-supervised and unsupervised approaches. IEEE Transactions on Knowledge and Data Engineering, 36(11):6124–6144.
Machado, M. R., Asadi, A., de Souza, R. W. R., and Ugulino, W. C. (2024). Green ai in the finance industry: Exploring the impact of feature engineering on the accuracy and computational time of machine learning models. Applied Soft Computing, 167:112343.
Medeiros, E. P., Machado, M. R., de Freitas, E. D. G., da Silva, D. S., and de Souza, R. W. R. (2024). Applications of machine learning algorithms to support covid-19 diagnosis using x-rays data information. Expert Systems With Applications, 238:122029.
Morteza, A., Yahyaeian, A. A., Mirzaeibonehkhater, M., Sadeghi, S., Mohaimeni, A., and Taheri, S. (2023). Deep learning hyperparameter optimization: Application to electricity and heat demand prediction for buildings. Energy and Buildings, 289:113036.
Riedmiller, M. and Lernen, A. (2014). Multi layer perceptron. Machine learning lab special lecture, University of Freiburg, 24:11–60.
Rigatti, S. J. (2017). Random forest. Journal of insurance medicine, 47(1):31–39.
Salehi, S. and Schmeink, A. (2024). Data-centric green artificial intelligence: A survey. IEEE Transactions on Artificial Intelligence, 5(5):1973–1989.
Shilong, Z. et al. (2021). Machine learning model for sales forecasting by using xgboost. In 2021 IEEE International Conference on Consumer Electronics and Computer Engineering (ICCECE), pages 480–483. IEEE.
Verdecchia, R., Cruz, L., Sallou, J., Lin, M., Wickenden, J., and Hotellier, E. (2022). Data-centric green ai an exploratory empirical study. In 2022 international conference on ICT for sustainability (ICT4S), pages 35–45. IEEE.
Verdecchia, R., Sallou, J., and Cruz, L. (2023). A systematic review of green ai. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 13(4):e1507.
Xu, J., Zhou, W., Fu, Z., Zhou, H., and Li, L. (2021). A survey on green deep learning. arXiv preprint arXiv:2111.05193.
Yu, M., Yang, H., Wei, X., Fu, S., Xu, J., and Zhang, J. (2026). Carbon footprint efficiency ratio: A unified indicator for green evaluation of evolutionary algorithms. IEEE Transactions on Emerging Topics in Computational Intelligence.
Zavyalova, E. B., Volokhina, V. A., Troyanskaya, M. A., and Dubova, Y. I. (2023). A humanistic model of corporate social responsibility in e-commerce with high-tech support in the artificial intelligence economy. Humanities and social sciences communications, 10(1):274.
Zhou, Y. and Bu, W. (2026). Artificial intelligence adoption, energy management, and corporate energy transition: Evidence from energy consumption, energy intensity, and carbon emission intensity. Energies, 19(3):821.
Introduction Related Work Methodology Datasets Feature Engineering and Feature Selection Experimental Design Machine Learning Models Carbon Monitoring Framework and Experimental Setup Evaluation Metrics Results and Discussion The Green AI Sweet Spot: Dimensionality Reduction Environmental Cost of Dimensionality Augmentation Statistical Validation Limitations and Future Work Conclusion
de Chavannes, L. H. P., Kongsbak, M. G. K., Rantzau, T., and Derczynski, L. (2021). Hyperparameter power impact in transformer language model training. In Proceedings of the second workshop on simple and efficient natural language processing, pages 96–118.
Delanoë, P., Tchuente, D., and Colin, G. (2023). Method and evaluations of the effective gain of artificial intelligence models for reducing co2 emissions. Journal of Environmental Management, 331:117261.
Georgiou, S., Kechagia, M., Sharma, T., Sarro, F., and Zou, Y. (2022). Green ai: Do deep learning frameworks have different costs? In Proceedings of the 44th International Conference on Software Engineering, pages 1082–1094.
Guo, G., Wang, H., Bell, D., Bi, Y., and Greer, K. (2003). Knn model-based approach in classification. In OTM Confederated International Conferences” On the Move to Meaningful Internet Systems”, pages 986–996. Springer.
Haakman, M., Cruz, L., Huijgens, H., and Van Deursen, A. (2021). Ai lifecycle models need to be revised: An exploratory study in fintech. Empirical Software Engineering, 26(5):95.
Jiao, Z., Wang, H., Xing, J., Yang, Q., Yang, M., Zhou, Y., and Zhao, J. (2023). A light-gbm based framework for lithium-ion battery remaining useful life prediction under driving conditions. IEEE Transactions on Industrial Informatics.
Kaur, R., Kundu, T., Park, K. M., and Pinsky, E. (2026). The carbon cost of intelligence: A domain-specific framework for measuring ai energy and emissions. Energies, 19(3):642.
Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., and Liu, T.-Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems, 30.
LaValley, M. P. (2008). Logistic regression. Circulation, 117(18):2395–2399.
Li, G., Yu, Z., Yang, K., Lin, M., and Chen, C. P. (2024). Exploring feature selection with limited labels: A comprehensive survey of semi-supervised and unsupervised approaches. IEEE Transactions on Knowledge and Data Engineering, 36(11):6124–6144.
Machado, M. R., Asadi, A., de Souza, R. W. R., and Ugulino, W. C. (2024). Green ai in the finance industry: Exploring the impact of feature engineering on the accuracy and computational time of machine learning models. Applied Soft Computing, 167:112343.
Medeiros, E. P., Machado, M. R., de Freitas, E. D. G., da Silva, D. S., and de Souza, R. W. R. (2024). Applications of machine learning algorithms to support covid-19 diagnosis using x-rays data information. Expert Systems With Applications, 238:122029.
Morteza, A., Yahyaeian, A. A., Mirzaeibonehkhater, M., Sadeghi, S., Mohaimeni, A., and Taheri, S. (2023). Deep learning hyperparameter optimization: Application to electricity and heat demand prediction for buildings. Energy and Buildings, 289:113036.
Riedmiller, M. and Lernen, A. (2014). Multi layer perceptron. Machine learning lab special lecture, University of Freiburg, 24:11–60.
Rigatti, S. J. (2017). Random forest. Journal of insurance medicine, 47(1):31–39.
Salehi, S. and Schmeink, A. (2024). Data-centric green artificial intelligence: A survey. IEEE Transactions on Artificial Intelligence, 5(5):1973–1989.
Shilong, Z. et al. (2021). Machine learning model for sales forecasting by using xgboost. In 2021 IEEE International Conference on Consumer Electronics and Computer Engineering (ICCECE), pages 480–483. IEEE.
Verdecchia, R., Cruz, L., Sallou, J., Lin, M., Wickenden, J., and Hotellier, E. (2022). Data-centric green ai an exploratory empirical study. In 2022 international conference on ICT for sustainability (ICT4S), pages 35–45. IEEE.
Verdecchia, R., Sallou, J., and Cruz, L. (2023). A systematic review of green ai. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 13(4):e1507.
Xu, J., Zhou, W., Fu, Z., Zhou, H., and Li, L. (2021). A survey on green deep learning. arXiv preprint arXiv:2111.05193.
Yu, M., Yang, H., Wei, X., Fu, S., Xu, J., and Zhang, J. (2026). Carbon footprint efficiency ratio: A unified indicator for green evaluation of evolutionary algorithms. IEEE Transactions on Emerging Topics in Computational Intelligence.
Zavyalova, E. B., Volokhina, V. A., Troyanskaya, M. A., and Dubova, Y. I. (2023). A humanistic model of corporate social responsibility in e-commerce with high-tech support in the artificial intelligence economy. Humanities and social sciences communications, 10(1):274.
Zhou, Y. and Bu, W. (2026). Artificial intelligence adoption, energy management, and corporate energy transition: Evidence from energy consumption, energy intensity, and carbon emission intensity. Energies, 19(3):821.
Introduction Related Work Methodology Datasets Feature Engineering and Feature Selection Experimental Design Machine Learning Models Carbon Monitoring Framework and Experimental Setup Evaluation Metrics Results and Discussion The Green AI Sweet Spot: Dimensionality Reduction Environmental Cost of Dimensionality Augmentation Statistical Validation Limitations and Future Work Conclusion
Publicado
19/10/2026
Como Citar
TEIXEIRA, Antonio Everton C.; SCARCELA, Fernanda; MARIANO, Lívian Torres; PEREIRA, Maria Clara; MACHADO, Marcos Roberto; FRANCO, Wellington; SILVA, Bruno R.; SOUZA, Renato William R. de.
Evaluating Computational and Carbon Costs of Machine Learning: The Impact of Feature Dimensionality During Model Development. In: ENCONTRO NACIONAL DE INTELIGÊNCIA ARTIFICIAL E COMPUTACIONAL (ENIAC), 23. , 2026, Cuiabá/MT.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 3131-3142.
ISSN 2763-9061.
DOI: https://doi.org/10.5753/eniac.2026.31177.
