Fourier Regression and ARIMAX Model for Forecasting Monthong Durian Price Index

The purposes of this study were to create models for forecasting the Monthong durian price index at the farm and to compare the forecasting efficiency of ARIMAX model (Autoregressive Integrated Moving Average with Exogenous Variable model) and Fourier Regression model. The data set was collected monthly between January 2014 and December 2020 (a total of 84 months). Production index and China consumer confidence were used as explanatory variables. The efficiency of both methods was compared by Mean Absolute Percentage Error (MAPE). From the result, we found that the ARIMAX (1,1,1) and Fourier regression models were both suitable for forecasting the Monthong durian price index. However, the MAPE value obtained from the ARIMAX model was 3.365 times higher than that obtained from Fourier regression model, suggesting that the Fourier regression model is more efficient. 

Keywords: Monthong durian price index; China consumer confidence; Fourier regression; ARIMAX model

*Correspondence author: Tel.: +66 86-4182542

                                                E-mail: kanittha_y@rmutt.ac.th