Exploring the effect of different number of periods in Thailand natural rubber supply forecasting models

Abstract

To dates, many studies attempted to forecast the price of natural rubber while many others also attempted to forecast the amount. The objective of this study was to explore the effect of different number of periods used in various forecasting models for the amount of natural rubber in Thailand. Monthly data from the Office of Agricultural Economics in the years 2007-2019 were obtained. Then the data were classified into 12 data sets; each of which had different number of periods including 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132 and 144 months. Decomposition forecasting methods consisted of the addition divide and the multiplicative divide methods were adopted. Model accuracy indicators were the mean absolute deviation, the mean square error and the mean absolute percentage error. The results showed that the amount of natural rubber and time were positively correlated at 99% confidence level. The most suitable model for forecasting the amount of natural rubber in Thailand found in our study was the multiplicative divide method by using the number of periods at 36 months yielding the mean absolute deviation of 66,264.61, the mean square error of 7,988,871,007.41 and the mean absolute percentage error of 24.66%.