APLICAÇÃO DE REDES NEURAIS ARTIFICIAIS NA PREVISÃO DO PREÇO DO MILHO NO ESTADO DO PARANÁ

Abstract

Abstract: Maize, a plant of Mesoamerican origin, is one of the most cultivated cereals in the world. Due to its nutritional capacity, it is widely used in human and animal food. The maize production chain is considered, in Brazilian agribusiness, as one of the most important. In this context, the present work aims to evaluate the effectiveness of models of Artificial Neural Networks (ANNs), for forecasting the price of maize in the state of Paraná. The database, made available by the Brazilian Agricultural Research Corporation (EMBRAPA), presents a historical series of the price of maize, in the state of Paraná, in the period between 2011 and 2020. Prediction models, based on BLSTM (Bidirectional Long Short-Term Memory) and MLP (Multilayer Perception) Neural Networks, were implemented in the Python language. Results obtained from the two models were compared. It was found, for a 6-month horizon, that the two forecast models provide reliable estimates for the price of maize.