19. Agriculture Commodity Prices Forecasting Using a Fuzzy Inference System

The objective of this chapter is to present a forecasting model of agricultural commodity prices using a Fuzzy Inference System. Recent studies have addressed the problem of commodity prices forecasting using different methods including artificial neural network and conventional model-based approaches. In this chapter, we proposed the use of a hybrid intelligent system called the Adaptive Neuro Fuzzy Inference System (ANFIS) to forecast agri-commodity prices. In ANFIS, both the learning capabilities of a neural network and reasoning capabilities of fuzzy logic are combined in order to provide enhanced forecasting capabilities compared to using a single methodology alone. Point accuracy of four agri-commodity prices (wheat, sugar, coffee, and cocoa) is appraised by computing root-mean-squared forecast errors and other well-known error measures. In terms of forecasting performance, it is clear from the empirical evidence that the ANFIS model outperforms over a feedforward neural network and two other conventional models (AR and ARMA).