A Hybrid Model for Improving Malaysian Gold Forecast Accuracy

Abstract

A hybrid model has been considered an effective way to improve forecast accuracy. This paper proposes the hybrid model of the linear autoregressive moving average (ARIMA) and the non-linear generalized autoregressive conditional heteroscedasticity (GARCH) in modeling and forecasting. Malaysian gold price is used to present the development of the hybrid model. The goodness of fit of the model is measured using Akaike information criteria (AIC) while the forecasting performance is assessed using bias, variance proportion, covariance proportion and mean absolute percentage error (MAPE).