A Comparison of Time Series Forecasting Methods for the Price of Batavia Pineapples Sent to Factories

Abstract

Forecasting using time series data analysis is an applicable method for predicting trends and seasonal variations in agricultural product prices. In this study, we aimed to create a model to predict the price of Batavia pineapples sent to factories. The data, which were obtained from the Office of Agricultural Economics from January 2007 to September 2023, were divided into training and test datasets. The time series of prices from January 2007 to December 2022 were used to create 8 forecasting models, including Box and Jenkins, Holt’s linear trend, Brown’s linear trend, damped trend, simple seasons, Winter’s additive, Winter’s multiplicative, and combined forecasting. The prices from January to September 2023 were then used to select a forecast model based on the root mean square error criterion. The selection results found that the combined Winter’s additive and multiplicative methods were the best, with a minimal forecasting error of only 0.26 baht/kilogram.