Forecasting the Price of Natural Rubber in Thailand Using the ARIMA Model

Abstract

Objectives of this research are to create the model that can be used to forecast natural rubber price of Thailand with independent variables such as world natural rubber price, synthetic rubber price, Advance market price of Tokyo Commodity Exchange (TOCOM), consumption of natural rubber, production of natural rubber with Autoregressive Integrated Moving Average (ARIMA) model. The study is divided as 4 procedures; first is time series analysis for forecasting natural rubber price of Thailand. It was found that the ARIMA(1,0,1) model is most suitable that can be explained for variations of natural rubber price of Thailand for 94.54% with Mean Absolute Percentage Error (MAPE) of 5.923. the second step is ARIMA analysis for forecasting natural rubber price of Thailand together with independent variables, it was found that ARIMAX(0,1,1) was most appropriate with variation explanation of Thailand’s natural rubber price of 99.89% and MAPE of 1.11. The third step is model analysis to forecast independent variables and the last step is using values from independent variable forecast in the third procedure to forecast Thailand's natural rubber price.