Building a Sustainable GARCH Model to Forecast Rubber Price: Modified Huber Weighting Function Approach

The unstable and uncertain nature of natural rubber prices makes them highly volatile prone to outliers, which can have a significant impact on both modeling forecasting. To tackle this issue, the author recommends hybrid model that combines autoregressive (AR) Generalized Autoregressive Conditional Heteroscedasticity (GARCH) models. utilizes Huber weighting function ensure forecast value remains sustainable even in presence outliers. study aims develop daily for 12-day period by analyzing 2683 price data from Standard Malaysian Rubber Grade 20 (SMR 20) Malaysia. analysis incorporates two dispersion measurements (IQR/3 Sn) three levels IO contamination 0%, 10%, 20%. results indicate using with IQR/3 measurement build AR(1)-GARCH(2,1) leads better sustainability. These findings potential enhance GARCH modifying M-estimator