Developed A Hybrid Sliding Window and GARCH Model for Forecasting of Crude Palm Oil Prices in Malaysia

Abstract

The increase of crude palm oil prices can significantly affect the worldwide economic activities. Therefore, an accurate model to forecast the crude palm oil prices is crucial so that necessary precautionary steps can be taken. In this study, a hybrid between Sliding Window and GARCH model was proposed to improve the forecasting accuracy of crude palm oil prices series. In this model, sliding window partitions is used to aggregate / cluster the original series into several number of constitutive series while GARCH model is utilized to forecast prices based on the selected window to complete variance calculation. A dataset of crude palm oil prices from Malaysian Palm Oil Board was used to test the performance of the proposed model. Direct application of GARCH model was used as a benchmark for effectiveness measurement with the proposed model by comparing mean percentage of absolute error and mean square error. The result has shown that the proposed hybrid sliding window and GARCH model demonstrates better forecasting performance than single GARCH model.