Análise preditiva da precificação de commodities pecuárias utilizando redes neurais LSTM

Abstract

This undergraduate thesis presents the development of an Artificial Neural Network (ANN) based on Time Series to make predictions on livestock commodity values, with a focus on beef. The system is capable of analyzing input data and generating graphs that show the expected predictions for the future values of these commodities, considering the prediction analysis of the price of the next day based on the previous ten days. Scenarios were carried out using time series with price data for a fat cow's arroba between 1997 and 2022, which were analyzed and discussed. The preliminary results of training and prediction demonstrated the effectiveness of the developed computational tool, regarding the accuracy of the implemented prediction model. Thus, the applicability of the LSTM Artificial Neural Network as an effective technique for predicting livestock commodity values, mainly beef, can be observed.