Prediction analysis of coal production of "b" company using ARIMA-asymmetric GARCH method

Coal has an important role in country growth. Coal production in Indonesia decreased 53.63 milion tons in 2020 because of Covid-19 Pandemic, declining coal consumption and low coal prices. Indonesia has many coal producers, one of them is "B"Company. "B"Company has not been able to meet its daily production target due to ups and downs in operation problem. So, predictive analysis of coal production is needed to help companies prepare for coal demand in the market. In this study, predictive analysis of coal production, especially actual coal getting "B"Company will be carried out in the daily production period in March 2021 - March 2022 using ARIMA-Asymmetric GARCH. The ARIMA method can be used on all data models, the data are not stationary, and have high accuracy for short-term predictions. Meanwhile the GARCH model can be used on data that has a heteroscedasticity effect on the residuals and Asymmetric GARCH can handle the asymmetric effects of the GARCH model. The goodness of the model and the level of accuracy will be seen through the values of AIC, SBC, RMSE and, MAD. The result of this study can also assist companies in making decision for preparation of coal production. Based on analysis, ARIMA([1,6],1,[1,13,14])-EGARCH(1,1) model is better than ARIMA([1,6],0,[1,13,14]) model, because it has an smaller AIC, SBC, dan RMSE values that is equal to 5,174.28, 5,188.82, and 2,500.98. Meanwhile, the MAD values for ARIMA([1,6],1,[1,13,14])-EGARCH(1,1) and ARIMA([1,6],0,[1,13,14]) models tend to be same that is equal to 2,381.58.