The Effectiveness of the Huber's Weight on Dispersion and Tuning Constant: A Simulation Study

Dispersion measurement and tuning constants are critical aspects of a model's robustness efficiency. However, in the presence outliers, standard deviation is not reliable measure dispersion Huber's weight. This research aimed to assess efficacy Huber weight function terms constant. The simulation study was conducted on hybrid autoregressive (AR) model generalized conditional heteroscedasticity (GARCH) with 10% 20% additive outlier contamination. In analysis, three measurements were compared: median absolute (MAD), interquartile range (IQR), IQR/3, two constant values (1.345 1.5). numerical results showed that during contamination IQR/3 outperformed MAD IQR. Our findings also potentially more robust 1.5 revealed decrease resistance outliers increased proposed value (h) AR(1)-GARCH(1,2) while minimising effect outliers.