Forecast of Prices of the Sugar using Artificial Neural Networks

Abstract

For the countries Latin American the production and export of agricultural products are of great importance in the balance of their commercial bills. Like this, to know the world commercial behavior of those products is necessary for the development of commercialization projects. As the participation of Brazil in the international market of sugar is elevated, approximately 40% of the exports world championships are national, the relationship of the brazilian domestic market of this commodity is expected with the international prices. Like this, this work will analyze the impact of the prices of the future contract of sugar rude nº11 of Csce/Nybot, commercial exchange rate, barrel of petroleum WTI Spot (North American market) and Europe Brent Spot (European market) in the prices of the granulated sugar conditioned in bags of 50 kg. In this analysis it will be used artificial neural networks (NN's) to explore the influence of the parameters: number of layers and numbers of intermediate neurons, on the acting of NN's. Keywords: Price-Forecast. Neural networks. Sugar. Alcohol.