Modelo de previsão da produção de ferro gusa no estado de Minas Gerais (2022 - 2024)

Abstract

The main goal of this study consists on study and applies some methods of data analyses that can be apply as tools, to support the decision-making for understanding and planning the pig iron market in Minas Gerais state. Thus, it focus is estimate a prediction model for the pig iron production in Minas Gerais between 2022 and 2024. To achieve this goal, a study of a group of statistical methods was applied to the data series, they are: exponential smoothing, Box & Jenkins and Dynamic Regression. To select the model to be used for the prediction, a competition was made between the methods previously mentioned through the analyses of out of-sample rolling evaluation. To support the estimation of the prediction model, a study was done about the pig iron sector in Brazil and Minas Gerais. The winner model indicates decrease of the Minas Gerais production of pig iron of 3.4% per year between 2022 and 2024.