A New Latex Price Forecasting Model to Reduce the Risk of Rubber Overproduction in Thailand

Abstract

One of the key areas in risk management in the public rubber industry in Thailand (PARIT) is to accurately forecast rubber latex prices thus to adjust rubber production in a timely manner. Accurately forecasting rubber latex price may not only reduce risks of overproduction and costs of over stocking, but also respond promptly and directly to global market thus improve in gaining higher sales in the competitive rubber marketing environment. This chapter presents a rubber latex price forecasting model, with three variations, i.e., one-year prediction, 6-month prediction and 4-month prediction, each embedding with either non-neural or neural network training techniques. The model is validated using actual rubber latex prices trend data, which in turn compared with experimental forecasting results to determine forecasting accuracy and the best-fitting model for policy makers in PARIT. © Springer-Verlag Berlin Heidelberg 2012.