Comparative analysis of techniques for forecasting u.s. dollar movements in the brazilian financial market: a Machine Learning approach

• Víctor Souza Santos UEFS
• Carlos Alberto Rodrigues UEFS

DOI: https://doi.org/10.5753/erbase.2025.13752

Abstract

This paper compares machine learning algorithms for predicting the dollar’s exchange rate against the real, seeking the most accurate model. Using historical and time series data, supervised classification techniques were applied. The following algorithms were evaluated: Random Forest, Neural Network, SVR, SVC, Linear Regression, and KNN. Random Forest and Neural Network stood out, with accuracies of 93.68% and 89.47%. SVC, SVR, and KNN performed poorly. Linear Regression performed intermediately, serving as a benchmark. The conclusion is that models capable of capturing non-linear relationships are more effective in forecasting exchange rates in volatile scenarios.

References

Breiman, L. (2001). Random forests. Machine Learning, 45(1):5–32.

Cortes, C. and Vapnik, V. (1995). Support-vector networks. Machine Learning, 20(3):273–297.

Haykin, S. (2001). Redes neurais: princípios e prática. Bookman, 2 edition.

Lin, X., Yang, Z., and Song, Y. (2011). Intelligent stock trading system based on improved technical analysis and echo state network. Expert Systems with Applications, 38(9):11347–11354.

Machado, E. J., de Assis, C. A. S., and Pereira, A. C. M. (2020). Modelagem, implementação e avaliação de estratégias de negociação baseadas em algoritmos de aprendizado de máquina para o mercado financeiro. Revista Brasileira de Computação Aplicada, 12(1):16–31.

Mali, K. (2024). Everything you need to know about linear regression. Blog post.

Ring, M. and Eskofier, B. M. (2016). An approximation of the gaussian rbf kernel for efficient classification with svms. Pattern Recognition Letters, 84:107–113.

Ryll, L. and Seidens, S. (2019). Evaluating the performance of machine learning algorithms in financial market forecasting: a comprehensive survey. arXiv preprint.

Singh, A. (2024). Knn algorithm: introduction to k-nearest neighbors algorithm for regression. Blog post.

Tsantekidis, A., Passalis, N., Tefas, A., Kanniainen, J., Gabbouj, M., and Iosifidis, A. (2017). Forecasting stock prices from the limit order book using convolutional neural networks. In Proceedings of the 2017 IEEE 19th Conference on Business Informatics (CBI), pages 7–12.