Forecasting State Revenues in Brazil Using Machine Learning: A Study on the Application of Predictive Models
State revenue forecasting, Machine Learning, Predictive models, Tax collection, Public policies.
This dissertation project aims to investigate the application of machine learning techniques for forecasting state revenues in the Brazilian context. Tax collection is essential for the fiscal management of states, and the accuracy in revenue forecasting plays a crucial role in the planning and execution of public policies. However, the complexity and volatility of factors influencing tax revenue pose significant challenges for traditional forecasting based on econometric models. In this regard, the use of machine learning models emerges as a promising approach, allowing the analysis of large volumes of data and the identification of nonlinear patterns. Through this study, we intend to explore different machine learning algorithms and techniques, such as neural networks, decision trees, and regression, using historical revenue data and other relevant economic variables. The expected outcome is the development of an accurate and reliable forecasting model that can assist state governments in making strategic decisions and improving fiscal management.