Forecasting Cassava Starch Price in Thailand by Using Time Series Models

Abstract

The purpose of this study is to forecast the cassava price in Thailand using five time series methods, which are Box-Jenkins method, Holt’s exponential smoothing method, damped trend exponential smoothing method, simple exponential smoothing method, and 3, 6, 12 month moving average methods. The secondary data obtained from Thai Tapioca Starch Association (TTSA) between January 2009 and December 2014 are used for time-series analysis. The data obtained from 2009 to 2013 were used to formulate the forecasting models and the data in 2014 were used to compare the performance of the forecasting methods via the criteria of the lowest mean absolute percentage error and root mean squared error. The findings from this research indicate that under the criterion of mean absolute percentage error, the 3 month moving average method is the most accurate to forecast the cassava price in 2014. In addition, under the criterion of root mean squared error, the simple exponential smoothing method is the most accurate. Additionally, the forecast values from both methods are reliable and no statistically significant difference was found between them.