Forecasting Model for the Export Values for Fresh Frozen Shrimp of Thailand

Abstract

The purpose of this research is to forecast the export values of fresh frozen shrimp using three-time series analysis methods including Box-Jenkins, Winters’ multiplicativeexponential smoothing, and combination forecasting. The secondary data obtained from the website of Office of Agricultural Economics with total 197 observations are used and divided into two series. The first 192 observations from January, 1998 until December, 2013 used to build the forecasting models and the last 5 observations from January until May, 2014 used to compare the forecasting methods performance via the criteria of the lowest mean absolute percentage
error and root mean squared error. Research findings indicated that for all forecasting methods that have been studied, Winters’ multiplicative exponential smoothing method is the most appropriate method for this time series.