A Comparison of MHW and AHW Methods for Forecasting Crude Palm Oil Productions in Thailand
Keywords:Time-series methods, Exponential smoothing, Seasonality, MAPE
A comparison of the Multiplicative Holt-Winters method (MHW) and the Additive Holt-Winters method (AHW) as the time-series methods for forecasting crude palm oil productions in Thailand is presented in this paper. Crude palm oil productions from January 2006 to September 2018 (i.e. 153 months) as the input data are collected from the database of the Department of Internal Trade, Ministry of Commerce, Thailand. The major contribution of our paper is that the well-known MHW and AHW methods, which are appropriately used for the input data with trend and seasonality behaviors, are tested and evaluated. Therefore, the best forecast results (i.e. the forecast crude palm oil productions) by the optimal method are determined. Our study demonstrates that the AHW method shows good results in forecasting crude palm oil productions. It provides the smallest forecasting error measured by Mean Absolute Percentage Error (MAPE). Forecast results of October 2018 to December 2018 (i.e. three months) and trends of the average monthly and yearly crude palm oil productions in Thailand from the past to the present are also reported and analyzed.