Intervention analysis measures the impact of various external events or interventions capable of changing data patterns. This research aims to determine the outliers handling on the seasonal ARIMA intervention model using the Box-Jenkins method. The pre-intervention model formed contains seasonal and step functions, which does not fulfill the white noise of the final intervention model. Therefore, the outliers need to be detected the model meets the white noise assumption. The intervention model and outlier detection in this study are conducted to capture the impact of a tariff-setting policy of 5 and 15 percent, called the first and second intervention, on the volume of Hot Rolled Coil/Plate (HRC/P) imports. When the outlier is detected, the next step is to examine and adjust its effect on the model by adding the effect of the outlier in the model. Using the seasonal ARIMA intervention model, the results showed that the first and second interventions significantly reduced the volume of HRC/P imports. A limitation of this research is that this model cannot include other independent variables in the modeling.
Copyrights © 2024