Selecting an Accurate Cacao Price Forecasting Model

Abstract

This study aims to analyze and select an accurate forecasting method for predicting cocoa prices. Monthly data of cacao prices at world and domestic markets from 2008:1 – 2016:12 are used in this research. Three univariate forecasting models are applied, i.e., ARIMA, Exponential Smoothing, and Decomposition. Selection of an accurate forecasting model use three accuracy indicators, namely, MAD, MAPE and MSD. The best cocoa price forecasting method is the ARIMA method. ARIMA is the most appropriate method used at both world and domestic prices.