Forecasting Agricultural Commodity Prices Using Artificial Neural Network (ANN)

<p>In view of increasing fluctuations and volatility on agricultural commodity markets, forecasting agricultural commodity prices became more relevant for management decisions. Obtaining an effective and accurate price forecasting will support decision maker towards a variety of decisions, such as storage decisions, hold and sell decisions and hedging decisions.</p>
<p>Theoretically, there is no possibility to reach better forecasts using forecasting methods in an efficient market because the observable price already reflects all available information and price s fluctuations in the future occur randomly. In reality, however, systematic patterns might be found in agricultural price series. Knowing and discovering these structures will facilitate the process of price forecasting.
Recently, Artificial Neural Networks (ANNs) are used to discover such structures. In this study, the ability of ANNs for forecasting agricultural commodity prices will be investigated. In order to clear some cloud in ANNs applications, a simulation experiment will be performed. Furthermore, different price series of crops and animal production will be used, whereas those products vary in seasonal fluctuations, volatility and possibility of storing. Out-of-sample technique will be used to evaluate the performance of forecasting. Both ARIMA models and commodity futures prices will be used as benchmarks.</p>