Comparison between Artificial Neural Network and ARIMA Model in Forecasting Palm Oil Price in Malaysia

Malaysia is one of the largest producers of palm oil in the world. Malaysia exported palm oil to almost 160 countries and contributed about RM50 billion or 3.5 per cent to Gross Domestic Product (GDP) in 2019. The palm oil industry is the fourth largest contributor to Malaysia's export earnings. However, palm oil price keeps fluctuating over time. Therefore, accurate prediction of palm oil price is important as investors deal with risks and uncertainties in the future. This study forecasts palm oil price in Malaysia using artificial neural network (ANN) and autoregressive integrated moving average (ARIMA) model. Comparisons between the two models are made and the most accurate model is selected. Monthly palm oil price data from January 2008 to December 2018 are used to build the forecasting models. The models are used to forecast the price of palm oil in Malaysia for year 2019. The predicted values are used to compare with the actual values. The main result reveals that artificial neural network model is more accurate compared to ARIMA model in forecasting the palm oil price although both models give good forecasting performance.