Hybrid of ARIMA and Quantile Regression (ARIMA-QR) Model for Forecasting Paddy Price in Indonesia

The price of paddy as the main food commodity in Indonesia, from year to year continues to experience fluctuations but tends to increase over the past few years. This requires that decision makers take action to maintain price stabilization. Indonesian Bureau of Logistics (BULOG) as a decision maker needs to know the forecasting of paddy prices over the next periods in order to determine the best actions or policies. The policy can be in the form of the amount of government paddy reserves, the amount of release of stock to the markef the determination of the amount of imported paddy and the price of paddy. In this research, the price of paddy is forecasted by using the ARIMA-QR method to obtain forecasting results for the future period as well as identifying factors that influence paddy price fluctuations. In doing this forecasting, several variables are used which influence the fluctuations in paddy prices such as the price of grain basis (GKG) and world paddy prices, the amount of BULOG stock, led holiday and the forecasting value of paddy prices that have been done previously. The data used is monthly data starting from 2000-2015. Based on the results of the study, the price forecasting model using ARIMA and ARIMA-QR has an accuracy of 1.47% for q = 0.25, 1.21% for q = 0.5 and 1.42% at the time q = 0.75. This average accuracy is 0.03% lower than the ARIMA accuracy.